Professional Documents
Culture Documents
Extended Conceptual Metaphor Theory
Extended Conceptual Metaphor Theory
Zoltán Kövecses
Eötvös Loránd University
University Printing House, Cambridge CB2 8BS, United Kingdom
One Liberty Plaza, 20th Floor, New York, NY 10006, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre,
New Delhi – 110025, India
79 Anson Road, #06–04/06, Singapore 079906
www.cambridge.org
Information on this title: www.cambridge.org/9781108490870
DOI: 10.1017/9781108859127
© Zoltán Kövecses 2020
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2020
Printed in the United Kingdom by TJ International Ltd, Padstow Cornwall
A catalogue record for this publication is available from the British Library.
Library of Congress Cataloging-in-Publication Data
Names: Kövecses, Zoltán, author.
Title: Extended conceptual metaphor theory / Zoltán Kövecses.
Description: Cambridge ; New York, NY : Cambridge University Press, 2020. |
Includes bibliographical references and index.
Identifiers: LCCN 2019040754 (print) | LCCN 2019040755 (ebook) |
ISBN 9781108490870 (hardback) | ISBN 9781108796620 (paperback) |
ISBN 9781108859127 (epub)
Subjects: LCSH: Metaphor. | Cognitive grammar.
Classification: LCC P301.5.M48 K67 2020 (print) | LCC P301.5.M48 (ebook) |
DDC 401/.43018–dc23
LC record available at https://lccn.loc.gov/2019040754
LC ebook record available at https://lccn.loc.gov/2019040755
ISBN 978-1-108-49087-0 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy
of URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
Very much belatedly. . ., to the memory of my
little sister, Annika.
Contents
References 186
Index 194
Figures
7.1 Summary of key ideas of the extended CMT view page 165
7.2 Four context types and their contextual factors 165
Preface
Acknowledgments
No book is written in a vacuum. In addition to the vast body of published work
on conceptual metaphor theory, in particular, and on metaphor, in general, a
large number of colleagues, anonymous reviewers of this book, fellow
researchers, students, conference participants, and just ordinary people inter-
ested in my work helped me with their ideas and suggestions in many ways.
I am thankful to them all. Special thanks go to the following people (in
alphabetical order): Kathleen Ahrens, Valentina Bambini, Antonio Barcelona,
Réka Benczes, Bogusław Bierwiaczonek, Anna Borghi, Mario Brdar, Rita
Brdar-Szabó, Szilvia Csábi, Alice Deignan, John Douthwaite, Rachel Giora,
Andrew Goatly, Patrick Colm Hogan, Robert Hoffman, Robert Kardela, Sonja
Kleinke, Francisco Ruiz de Mendoza, Andreas Musolff, Uwe-Klaus Panther,
Frank Polzenhagen, Michele Prandi, Günter Radden, Elena Semino, Veronika
Szelid, Elzbieta Tabakowska, John Taylor, Linda Thornburg, and Cliff
Winters. I am grateful to all of them. Several of my students in Budapest
and Heidelberg have helped me shape the ideas in this book: Mohsen Bakhtiar,
Olga Boryslavska, Orsolya Farkas, Eszter Nucz, Orsolya Putz, and Erzsébet
Tóth-Czifra. Olga Boryslavska helped me with several of the figures in the
book. I am thankful to them for the many, sometimes passionate but always
enjoyable, discussions. I am grateful to Helen Barton at CUP for her encour-
agement and support.
Two people have played very special roles in the course of my career as a
metaphor researcher. Ray Gibbs has always been available to discuss various
issues related to the field and he and his work gave me a huge amount of
encouragement and inspiration. And last but definitely not least, without
George Lakoff I could not have and would not have done any of my work
on metaphor.
1 A Brief Outline of “Standard” Conceptual
Metaphor Theory and Some Outstanding Issues
Conceptual metaphor theory (CMT) started with George Lakoff and Mark
Johnson’s book, Metaphors We Live By (1980).1 The theory goes back a long
way and builds on centuries of scholarship that takes metaphor not simply as
an ornamental device in language but as a conceptual tool for structuring,
restructuring, and even creating reality. Notable philosophers in this history
include, for instance, Friedrich Nietzsche and, and more recently, Max Black.
A recent overview of theories of metaphor can be found in Gibbs (2008) and
that of CMT in particular in Kövecses (2002/2010).
Since the publication of Lakoff and Johnson’s (1980) work, a large amount
of research has been conducted that has confirmed, added to and also modified
their original ideas. Often, the sources of the new ideas were Lakoff and
Johnson themselves. Given this situation, it is obvious that what we know as
CMT today is not equivalent to the theory of metaphor proposed in Metaphors
We Live By. Many of the critics of CMT assume, incorrectly, that CMT equals
Metaphors We Live By. For this reason, I will not deal with this kind of
criticism in this introduction to CMT.
The standard definition of conceptual metaphors is this: A conceptual meta-
phor is understanding one domain of experience (that is typically abstract) in
terms of another (that is typically concrete) (see Lakoff and Johnson 1980).
This definition captures conceptual metaphors both as a process and a product.
The cognitive process of understanding a domain is the process aspect of
metaphor, while the resulting conceptual pattern is the product aspect. In this
survey of the theory, I will not distinguish between the two aspects.
In the following sections, I attempt to identify and briefly describe the main
features of CMT, as I see them. Other researchers might emphasize different
properties of the theory. At the same time, I tried to select those features
on which there is some agreement among practitioners of CMT. At the end of
the chapter and given the description of the properties, I list a number of
1
This chapter is a revised and enlarged version of a chapter I wrote entitled “Conceptual metaphor
theory” in Elena Semino and Zsófia Demjén, eds., 2017, The Routledge Handbook of Metaphor
and Language, 13–27. London: Routledge.
1
2 A Brief Outline of “Standard” Conceptual Metaphor Theory
outstanding issues in CMT – issues that we need to get clear about in order to
make CMT an even more powerful theory of metaphor.
Charteris-Black (2017) points out, there can be subtle differences between, for
instance, the causes and intensities of target concepts that are indicated by
various specific lexical items. As an example, we can mention the words
kindle, spark, inflammatory, and so on, that refer to different kinds of causes.
The mappings as usually formulated are not sensitive to this level of
specificity. (But the way this can be remedied in CMT will be discussed in
Chapters 4–6).
The generic set of mappings is systematic in the sense that it captures a
coherent view of fire that is mapped onto anger: There is a thing that is not
burning. An event happens (cause of fire) that causes the fire to come into
existence. Now the thing is burning. The fire can burn at various degrees of
intensity.
Similarly for anger: There is a person who is not angry. An event happens
that causes the person to become angry. The person is now in the state of
anger. The intensity of the anger is variable.
The mappings bring into correspondence the elements and the relations
between the elements in the fire domain (source) with elements and the
relations between the elements in the anger domain (target). Indeed, it seems
reasonable to suggest that, in a sense, the mappings from the fire domain
actually bring about or create a particular conception of anger relative to the
view of fire we have just seen. This is what it means that a particular source
domain is used to conceptualize a particular target domain. (I will come back
to this issue later.)
In many cases, however, the two-domain account does not work and must
be supplemented by a model of explanation that relies on four (or more)
domains, or spaces (see Chapter 6 on conceptual integration and metaphor).
metaphor, there are many things that cannot be mapped, or carried over, from
the source to the target. For example, given that theories are buildings,
the number of rooms or whether the building has a cellar or an attic are not
mapped. Several explanations have been offered to delimit the amount of
knowledge that can be transferred from the source. One of them is the
“invariance hypothesis” developed by Lakoff (1990). It suggests that every-
thing from the source can be mapped onto the target that does not conflict with
the image-schematic structure of the target. Another is proposed by Grady
(1997a,b), who claims, in essence, that those parts of the source domain can be
mapped that are based on “primary metaphors” (on “primary metaphors,” see
Grady (1997a,b)). Finally, Kövecses (2000a, 2002/2010) proposed that the
source maps conceptual materials that belong to its main meaning focus or
foci. It should be noted that the three suggestions differ with respect to which
part of a conceptual metaphor they rely on in their predictions concerning what
is mapped. The first one relies primarily on the target, the second on the
connection between source and target, and the third on properties of the source.
None of these are entirely satisfactory.
our conceptual system governs how we act in the world, we often act
metaphorically.
When we conceptualize an intangible or less tangible domain metaphoric-
ally as, and from the perspective of, a more tangible domain, we create a
certain metaphorical reality. We imagine life one way when we think of it as a
journey (see the aforementioned discussion), and in another way when we
think of it as a theatre play, as reflected in Shakespeare’s famous lines “All the
world is a stage / and all men and women are merely players.” The two source
domains result in very different views on life, and in this sense they create very
different realities.
Whenever a new source domain is applied to a particular target, we see the
target domain differently than we saw it before. The limiting case of this
situation is the one when a particular target domain does not exist at all, but
by the application of one (or several) source domain(s), it actually gets created.
Very often, the etymologies of words for abstract concepts reflect this early
conceptualization. For example, comprehension (“understanding”) is
clearly an abstract concept. Given the understanding is grasping con-
ceptual metaphor (as in “I did not grasp what he said,” “He is slow on the
uptake”), it makes sense that the English word comprehend derives from the
word that means “grasp” in Latin.
This kind of “reality construction” is very common in advertising, where,
often, interesting or amusing cases of metaphorical reality get created. When
advertisements for, say, deodorants promise “24-hour protection,” they make
us see a deodorant as our helper or ally in a fight or war against an enemy. The
enemy is no other than our own body odor. So if we did not think of our body
odor as our enemy before, i.e., as something we have to be protected against,
the advertisements can easily make us view it as such. In this manner, the
metaphors used in advertisements and elsewhere can create new realities for
us. Such realities are of course metaphorically defined. But this does not make
them unimportant for the way we live. If we think of our body odor as
something we need to be protected against and as a result go and buy a
deodorant to overcome the enemy, we are clearly thinking and acting
according to a metaphorically defined reality. This is a further example of
how the implications of a source domain for a particular target can be utilized
(in a process I called metaphorical inference or entailment in the previous
section).
and Turner (e.g., Fauconnier and Turner 2002) in their discussion of concep-
tual integration, where they assume a “network model” that consists of
four spaces (see Chapter 6 for details). Richard Trim (2007, 2011) uses the
term “metaphor network” for the development of an individual metaphor,
such as the lion as a symbol of courage in British heraldry, as well as for a
network of metaphors that grows historically from a core concept, such as
mother or father – giving rise to such metaphors as mother tongue or
mother country in English. He shows how such networks emerge in a number
of European languages and cultures over the ages. Needless to say, both of
these are extremely interesting and important applications of the idea of
“network,” but I have an application of the term in mind that is different
from both.
As we have seen above, the source domains of conceptual metaphors
constitute coherent organizations of experience, and the mappings from the
source onto the target domains create equally systematically organized target
domains. But the question is whether such systematic source to target map-
pings are isolated from each other. I suggest that they are likely to belong to
larger, hierarchically organized systems of metaphors.
The principles for the organization of such metaphor systems can be of
several distinct kinds. In one (a), the metaphors are organized in a straightfor-
ward hierarchy such that both the source and the target are specific cases of
higher generic-level concepts. In another (b), different aspects of a given
generic-level concept can be differentially conceptualized by means of con-
ceptual metaphors. In still other cases (c), a single aspect of several different
abstract concepts may organize a large number of subordinated specific-level
conceptual metaphors into a hierarchy. In a fourth (d), the conceptual meta-
phors form a system because the target domains are part of an independently
existing hierarchy of concepts. In a fifth (e), what connects the conceptual
metaphors and makes them form a system is the fact that a particular specific
level target concept is a special case of a number of different higher-level
concepts that have their own characteristic conceptual metaphors. There are
probably additional ways in which metaphor systems (or networks) are
formed, but for the present purposes it is sufficient to take these five possibil-
ities into account and briefly describe them.
Straightforward Hierarchies
In this case, both the source and the target are specific-level concepts of
generic-level conceptual metaphors. This is the simplest and most straightfor-
ward type of hierarchy, and it involves a large number of cases. Let us take the
well-known anger is a hot fluid metaphor. This is an instance of the
generic-level metaphor emotions are forces. Actually, the hot fluid
1.15 Metaphor Networks 13
source can be further specified, yielding, for example, the concept of stew as
a potential source domain. We can represent this as follows:
emotions are forces
anger is a hot fluid in a container (He was boiling
with anger.)
anger is a stew (He was stewing.)
We can find the same situation for love:
emotions are forces
love is a natural force (I was overcome with love.)
love is the wind (It was a whirlwind romance.)
As we will see throughout this book, despite, or rather, because of, its simplicity,
this kind of network is crucially important for the extended theory of metaphor
I am going to present in this work. I will come back to it in several later chapters.
The conceptual metaphors for friendship emerge from these various metaphor
systems. Specifically, we find metaphors such as the following in the descrip-
tions of friendship:
State metaphor system:
friendship is a possessed object (My friendship with her did not
last long.)
Communication metaphor system:
sharing (communicating) experiences is sharing objects:
(We share intimate things with each other.) (Kövecses 1995a, 2000b)
The metaphor arises because communication between friends often involves
sharing ideas and feelings.
Interaction metaphor system:
interactions in friendship are economic exchanges (There is a
lot of give and take in our friendship.) (Kövecses 1995a)
16 A Brief Outline of “Standard” Conceptual Metaphor Theory
But the most commonly and strongly expressed criticism concerns meth-
odological issues; namely, how to identify metaphors in discourse, how the
study of metaphor should be based on real data (rather than just lexical or
intuitive data), and others (see, e.g., Deignan 2005; Pragglejaz Group 2007).
As I indicated above, we should now take these developments as an integral
part of CMT. However, the issue of the need to use real data for metaphor
analysis reveals an apparently real weakness of CMT: it is that CMT
researchers do not pay sufficient attention to the discourse and social-
pragmatic functions of metaphor in real discourse. This sounds like a valid
point. However, I do not think that CMT should be thought of as a view of
metaphor whose only job is to collect metaphorical expressions, set up
conceptual metaphors based on the expressions, lay out the mappings that
constitute those conceptual metaphors, and see how the particular conceptual
metaphors form larger systematic groups. A large further part of the mission
of CMT is to describe the particular syntactic, discursive, social, pragmatic,
rhetorical, aesthetic, etc., behavior and function of the metaphors in real data.
And this is precisely something that is currently conducted by a great number
of researchers (e.g., Low et al. 2010). But, to my mind, these researchers
are not competing with more “traditional” CMT scholars; instead, they are
working out an aspect of CMT that was “neglected” by CMT scholars. The
addition is necessary and more than welcome. This kind of work is just as
much part of CMT as other aspects of the theory. In other words, I find that
the “neglect” was not really neglect. The lack of sufficient attention to the
syntactic, pragmatic, etc., features of metaphors resulted from CMT scholars’
effort to add a cognitive dimension to metaphor that was mostly lacking in
previous work. This was, and still is, a major part of CMT scholars’ mission,
in collaboration with other metaphor researchers. Without pursuing that
mission, we would not have been able to link the study of metaphor with
the study of cognition in general.
(e) Several metaphor networks have been noted above. Are all of them
equally significant? Does any one of them stand out as especially import-
ant for the overall description of the metaphorical conceptual system?
(f ) How can the recently proposed view of deliberate metaphors be incorpor-
ated into a CMT framework? Or does it present a challenge to CMT?
(g) How can we account for the socio-pragmatic function of metaphorical
linguistic expressions in naturally occurring discourse? What would a
theory of conceptual metaphors look like that is capable of explaining the
conceptual structure of metaphors, as well as the discourse functions of
metaphorical expressions realizing them within a unified framework?
(h) Indeed, in connection with the previous question, should we think of
metaphor as a cognitive process that is happening online or as a product
that is somehow a part of a more or less stable conceptual system in long-
term memory?
(i) Though it is true that theories of metaphor comprehension make exten-
sive use of the notion of context, theories of metaphor production within
CMT have almost completely ignored the notion. Is it possible to propose
a theory of conceptual metaphors within the cognitive linguistic para-
digm that would integrate the context of metaphor production into CMT?
(j) The major idea of CMT is embodiment on which primary metaphors are
based. While clearly crucial, one can legitimately raise questions about
the exclusivity of embodiment in CMT. Isn’t it conceivable that there are
additional factors that play an important role in the creation of meta-
phors – both conceptual and linguistic?
(k) Concerning metaphorical creativity, CMT offers sophisticated accounts
of the novel metaphors in both every day and literary discourse. How-
ever, these accounts of metaphorical creativity have their limitations:
they can only explain creativity that is the result of putting together
potentially universal primary metaphors in unique ways or as a result
of a small number of conceptual devices (extending, elaboration, ques-
tioning, combining). However, in many cases we find novel metaphors
(both conceptual and linguistic) that require us to take into consideration
a variety of contextual factors, and not just universal body-based meta-
phors or universal cognitive processes.
(l) If we think with the help of conceptual metaphors, as we like to claim, it
is reasonable to ask why we use mixed metaphors in the actual produc-
tion of discourse. Wouldn’t it be easier and more efficient to use the same
conceptual metaphor throughout a given discourse? As a matter of fact,
metaphorically homogeneous discourse seems to be much rarer than
metaphorically heterogeneous discourse (i.e., mixing metaphors in dis-
course). Why is this the case? Can we provide a coherent CMT account
of mixed metaphor in discourse?
1.18 Extending CMT 21
(m) Cognitive linguists (including the present author) like to think of meta-
phors as “conceptual metaphors,” ignoring the diversity of phenomena
we call metaphor. What other kinds of metaphor are there, how can they
be characterized, and what is their relationship to conceptual metaphors?
Undoubtedly, there are many more questions that could be raised concerning
CMT. However, I find the questions in a. through m., among possibly others,
inevitable to raise in order to produce an extended, more comprehensive, and
hopefully improved, version of CMT. I address these issues in the following
chapters of the book.
2 The Abstract Understood Figuratively, the
Concrete Understood Literally, but the
Concrete Understood Figuratively?
1
The chapter reuses materials from previously published work in a revised form. In particular:
Kövecses, Zoltán. 2017a. A radical view of the literal-figurative distinction. In A. Benedek and
Á. Veszelszki, eds., Virtual Reality – Real visuality. Virtual, Visual, Veridical. 17–28. Frankfurt
am Main: Peter Lang.
22
2.1 Delimiting the Literal 23
handles the distinction between literal and metaphorical in the following way:
“those concepts that are not comprehended via conceptual metaphor might be
called ‘literal.’” But this still leaves us with cases that are hard to classify. For
example, it is clear that “Time passed quickly” and “The time will come” are
metaphorical due to the time is motion conceptual metaphor. But what
about early in sentences like “March comes earlier than April,” a sentence used
in one of Boroditsky’s (2001) experiments? Is early literal or metaphorical? It
is usually taken to be literal. But in the experiment Boroditsky found that
speakers of English responded faster (saying TRUE/FALSE) when they
received a horizontal prime, and speakers of Chinese were faster when they
received a vertical prime. The results of this priming study indicate that a word
we would take to be literal (since it is unrelated to motion) is comprehended
metaphorically (since speakers are affected by a conceptual metaphor – hori-
zontal versus vertical motion – in processing it). Thus, if the “comprehension
via conceptual metaphor” criterion is correct, the word would have to be taken
as metaphorical, not literal.
The criterion of “comprehension via conceptual metaphor” would give us a
similar dilemma in a large number of cases, including that of the domain of
time. As a matter of fact, the word time is open to the same question: is it literal
or metaphorical? If the word early is understood metaphorically, and thus
counts as metaphorical, then surely the word time is also understood meta-
phorically and has thus a metaphorical meaning. Or should we distinguish the
name (the word) of a domain from the domain itself (the concept of time)?
But can we separate word meaning from conceptualization in a cognitive
linguistic framework? Isn’t word meaning dependent of conceptualization
(see Langacker 1987)?
We encounter the same problem in all the cases where there is an abstract
target domain (anger, theory, time, desire, etc.) involving many lin-
guistic expressions that we would not hesitate to take to be literal (such as
anger, rage, ire, irritation, fury for anger; theory, idea, scheme, belief for
theory; time, period, moment, minute, hour for time; desire, urge, longing,
want, yearn for desire; etc.). Can abstract concepts such as these be literal?
That’s how we feel about them intuitively. But this should not be the case if we
apply the criterion of “comprehension via conceptual metaphor.” Clearly, the
concepts corresponding to them are all comprehended metaphorically via a
number of different conceptual metaphors, such as anger is fire, know-
ledge is vision, time is motion, desire is hunger, etc. My claim here
would be that abstractions can only be conceptualized metaphorically, and
consequently the linguistic items expressing these metaphorical conceptual-
izations can only be metaphorical.
However, we do not view them as such. The question is why. It seems to me
that in such cases we apply a criterion different from “comprehension via
2.1 Delimiting the Literal 25
literal (1) dead metaphors, (2) abstract entities or events without explicit
morphological indicators or markers for particular source domains, and (3)
the large class of cases where a concrete entity or event is figuratively
conceptualized historically, we are left with a rather limited set of cases.
Actually, it is very difficult to assess the exact size of this remaining set (but
see, e.g., work by Borghi et al. 2017; Altarriba et al. 2004). Based on research
in a variety of disciplines, we can probably include in that set basic-level
objects and events, a variety of spatial relations, basic emotions, body parts,
and probably some others (see Lakoff 1993).
But can we? Even here, we run into problems. In all of these groups, we find
items that are comprehended/constructed figuratively. For example, the basic-
level category of wolf in Hungarian is farkas, literally meaning “the one with a
tail,” or “that which has a tail.” This is the result of comprehending wolves
metonymically due to some taboos on naming this animal in the ancient belief
system. The English preposition in corresponds to the suffix –ban/ben in
Hungarian, and the Hungarian preposition goes back to the word bél
(“bowels”) in its etymology. In this case, the body serves as the source domain
for spatial relations. The English basic emotion term anger, as we saw above,
probably has a metonymic origin. Surprise (deriving from old French, and
ultimately Latin, surprendre), another basic emotion term, is the product of a
metaphorical process that can be characterized by the metaphor surprising
(someone) is unexpected seizure/attack (see Kövecses 2015b).
Finally, the French word for the face, visage, derives from the Latin visus
“sight.” This is probably also metonymic. What the examples show is that
even concepts that we can take to be based on our most basic and direct
experiences, such as the animals that surround us in nature, fundamental
spatial relations, basic emotions, and the parts of the human body can all be
conceptualized figuratively.
In other words, under a broad definition of what is figurative, very little of the
literal remains. This could seriously undermine a major claim of conceptual
metaphor theory; namely, that a sufficiently large part of language and thought
is about concrete things and is thus literal, and that we use this language and
thought to conceptualize abstractions. What we saw above was that our most
concrete experiences can be conceptualized figuratively, and not literally. But
this does not mean that they cannot be used as source domains for figurative
conceptualization. Rather, we seem to have a huge body of figuratively con-
ceptualized concrete experiences that we take over from previous generations,
use them as if they were entirely literal experiences, and turn them into vehicles
for the further conceptualization of more abstract things. Ultimately, much of
what we take to be literal is only ontologically literal (i.e., constitutes basic
concrete experience), but as regards its cognitive status it is also figuratively
construed/constructed experience. We reuse this apparently literal conceptual
2.3 Smell as a Target Domain 29
Existence
Let us begin with existence. Consider some examples for how the existence of
smell is presented by a dictionary:
30 Abstract Figuratively, Concrete Literally, Concrete Figuratively?
VERB + SMELL be filled with, have The air was filled with a pervasive smell of
chemicals. The cottage had a musty smell after being shut up over the winter. | give off
The skunk gives off an unpleasant smell when attacked. | catch, detect As she walked
into the house she detected the smell of gas. (boldface in the original)
The examples indicate a variety of ways in which the existence of smell can be
conceptualized. In the case of fill, the smell is viewed as a substance and the
object or location with the smell is viewed as a container, and the existence of
the smell as the substance being in the container. In the case of have, existence
of the smell corresponds to the possession of an object that is smell itself. In
the case of give off, causing smell to come into existence is the transfer of an
object and the source of the smell is a surface. Finally, catch and detect are not
dedicated to denoting the existence of smell, they simply assume its existence.
The examples and their construals of the existence of smell they reveal point
to the following conceptual metaphors:
filled with:
smell is a substance
the object / location that has the smell is a container
the existence of smell is for the substance to be in the container
have:
smell is an object
the existence of smell is the possession of an object
give off:
smell is an object
causing smell is transferring the object
source of smell is a surface
Intensity
Another generic-level concept that is linked to smell is intensity. In the case of
smell, intensity involves a scale of values from weak to strong. Collocation
dictionaries give us a good idea of this notion in relation to smell:
2.3 Smell as a Target Domain 31
These conceptual metaphors that relate to the intensity aspect of smell also
have their generic-level variants:
overpowering, strong, faint: intensity is strength
sharp, pungent: intensity is sharpness
vaguely: intensity is brightness
pervasive, fill: intensity is quantity
Lack of Control
The issue of control is also relevant to smell and perception in general.
A curious property of perceptual experience is that we do not have control
over what we perceive. If our sense organs are normal and no external
circumstance blocks the perception, we undergo the perception of any existing
perceptual stimulus. This is also often rendered in metaphorical ways – one of
the chief ways among them (according to dictionary information) is by means
of the verb hit in the case of smell. Consider the example taken from a
dictionary: Then the pungent smell hit us - rotting fish and seaweed.
We can analyze the metaphorical use of hit as follows:
smell is a physical force
intense smell is a strong physical force
sensing an intense smell is coming into contact with a
strong physical force
32 Abstract Figuratively, Concrete Literally, Concrete Figuratively?
The three basic conceptual metaphors above can be combined into a complex
one that describes the metaphorical conceptualization of the perceptual experi-
ence of strong smell:
the experiencer of an intense smell is a person coming into
abrupt physical contact with a strong physical force
But most importantly, the main point is, as before, that the specific-level
conceptual metaphors are licensed by corresponding generic-level ones:
causes are forces
intensity is strength
sensing is contact
Again, these generic-level metaphors pervade the conceptual system and allow
us to conceptualize various kinds of causes, intensity, and sensation, not just
smell as a cause, not just the intensity of smell, and not just sensing smell. In
other words, smell as an extremely basic and direct experience has aspects to it
that lend themselves to metaphorical conceptualization.
2.4 Conclusions
On the basis of the evidence we saw in this chapter, it seems that figurativity
(i.e., understanding concepts by means of metaphor or metonymy) does not
really distinguish concrete concepts from abstract ones. Both types of concepts
(concrete and abstract) are characterized by being understood metaphorically
and metonymically. As a result, we face a dilemma in accounting for the
unidirectionality of metaphorical mappings in conceptual metaphors. We
cannot claim that the concrete is understood in a literal way and we make
use of this in comprehending and creating abstractions.
Does this situation mean we have to discard CMT? In my view, it should not
mean that. We can propose, given the evidence we saw, that concepts consist
of two parts: an ontological part that specifies what kind of ontological
materials make up a concept and a cognitive part that specifies the ways in
which concepts are cognitively construed or constructed. I assume that this
distinction is similar, but at least related, to a distinction made by Langacker
(1987, 2008) between the content part and the construal part of concepts.
Now, given the distinction, we can suggest that in the case of literal meaning
for concrete concepts, the ontological content predominates over the cognitive-
construal part. In many instances, the construal part is limited to certain
generic-level dimensions of concepts, as we saw in the case of smell. These
dimensions that are shared across many concepts in the conceptual system are
figuratively (metaphorically, metonymically) understood. By contrast, in the
case of abstractions, the construal part predominates over the ontological part.
2.4 Conclusions 33
1
This chapter is a revised and enlarged version of Kövecses (2013).
34
3.1 A Brief Note on the Conceptual System 35
Lakoff and Johnson 1980, 1999), metonymy appears to be more clearly and
importantly connected with metaphor, as argued by several researchers, such
as Barcelona (2000); Kövecses (2002/2010); Kövecses and Radden (1998);
Lakoff and Kövecses (1987); Radden (2002); Radden and Kövecses (1999);
Taylor (1989), and others (in, e.g., the volume edited by Dirven and Pörings
2002). In contrast, several recent influential researchers within the conceptual
metaphor paradigm believe that correlation metaphors (best known as
“primary metaphors”) come into existence independently of, or without any
recourse to, metonymy (most notably, Grady 1997a,b; Grady and Johnson
2002; Lakoff and Johnson 1999).
As will be clear from the chapter, I align myself with the former group and
suggest that many metaphors (of the correlational kind) derive from metonymies,
that is, they have a metonymic basis. What distinguishes my position from the
view of the other proponents in the group that favors a metonymy-based
emergence for many metaphors is that I attempt to establish the relationship
between metaphor and metonymy by relying on several particular characteristics
of the conceptual system, as we know it today.
My specific proposal in the chapter will be that correlation-based metaphors
emerge from frame-like mental representations through a metonymic stage.
I suggest this happens when one of the elements of a frame-like mental
structure is generalized (schematized) to a concept that lies outside the initial
frame in a different part of the conceptual system. The generalization process
leads to sufficient conceptual distance between the initial and the new frame on
which metaphors can be based.
There are two kinds of structure that characterize the conceptual system.
Conceptual systems are organized “vertically,” which, essentially, provides for
a thematic structure in the system (see, e.g., Rosch 1978) and also “horizon-
tally,” which, essentially, defines individual concepts and consists of smaller
domains, or frames, or idealized cognitive models (see, e.g., Barsalou 1999;
Fillmore 1982; Lakoff 1987; Langacker 1987).
and distinct from them (i.e., is outside the sadness frame), it can now be seen
as a (source domain of a) metaphor for sadness: hence the conceptual metaphor
sad is down. sadness is in the emotion-dominated part (target) of the
hierarchy, whereas down(ward) is in the (more concrete) space-dominated
part (source).
The conceptual metaphor sad is down resulting from the generalization,
or schematization, of various behavioral responses motivates, or, to use
different cognitive linguistic terminology, licenses, or sanctions, a number
of metaphorical linguistic expressions, such as “to feel down,” “to be in low
spirits”, and the like. But, possibly based on this generalization, we can
develop further metaphorical conceptualizations in the following way: The
notion of down can be elaborated by a large number of more specific
instances of space that have nothing to do with sadness as conceived in the
sadness frame, including places with a downward orientation, such as pits
(resulting in “be in the pits”), dumps (resulting in “be down in the dumps”),
and the like. This is the second way in which downward bodily orientation can
lead to metaphoric conceptualization.
Thus, given the example above, we can suggest that metaphor appears to be
based on and derive from metonymy (as suggested by Lakoff and Kövecses for
anger and by several other authors, e.g., Barcelona 2000; Radden 2002). The
details of this development in many cases, such as the sad is down metaphor,
involve the process of generalization, or schematization, from specific bodily
responses characterized by a downward orientation (leading to a more general
spatial concept, down, in the case of sadness) and (then) the process of
specialization, or elaboration (leading to elaborations of the same schematic
spatial concept). Kövecses and Radden (1998) make a similar argument for the
source domain of heat in relation to anger in the conceptual metaphor anger
is heat (Lakoff and Kövecses 1987).
Consider now another example – the relationship that obtains between
closeness and intimacy. This is usually conceptualized, and viewed, as the
metaphor intimacy is closeness. But, similar to the previous cases, we can
also conceptualize the relationship between the two concepts as metonymy.
Since children and adults in general who are in an intimate relationship with
each other tend to be physically close (and, as a result, be able to touch each
other, feel the body temperature of the other, be able to smell the other, etc.),
we can think of it as a metonymy: closeness for intimacy. However, the
physical closeness that characterizes intimacy is specifically between two
people – with the several specific consequences just mentioned. Thus the less
generic version of the metonymy is physical closeness between two
people for intimacy. Similar to the generalization process we have seen
for sadness above, the physical closeness between two people is
3.2 Metaphor or Metonymy? 39
have access to the whole frame (guarding) – a situation that yields the meton-
ymy visual monitoring for guarding (or other kinds of control).
Visual monitoring is a kind of seeing, just as guarding and other forms of
keeping control are kinds of control. Given the initial frame in which they
are correlated, when visual monitoring is generalized (schematized) as seeing
and when various forms of control, such as guarding, are generalized
(schematized) as the notion of control in very different vertical hierarchies
of the conceptual system (control vs. vision), we conceptualize control
through seeing. We can think of this relationship as a metaphor, rather
than metonymy, because seeing and control are in distant parts of the
conceptual system (perception vs. social relation, or the like). But, in all
probability, they are brought together in the metaphor in the first place because
their more specific versions (visual monitoring and guarding) are both parts of
the same (initial) frame, or functional domain (or scene). Without this
co-occurrence in the same frame (yielding a metonymy), we would probably
not have thought of conceptualizing the control domain as seeing, yielding
a metaphor.
Given this analysis, the following general picture seems to emerge: One of
the events or states within a single initial frame (the original frame that
becomes the metaphorical target domain, such as intimacy or guarding/
control) gives rise to a distinct frame that becomes the metaphorical source.
We can represent this as follows:
metaphor
metaphor
However, in both cases, given two events or states in an initial frame, one
event or state stands metonymically for the other event or state within the same
frame, and, then, one of the events or states is generalized (schematized) in
such a way that it becomes independent of the original frame. When the latter
happens, we find the two initially co-occurring and correlated events or states
within the same frame emerging in two distinct and conceptually distant
frames in different parts of the conceptual system. We can now think of such
cases as metaphoric connections between the frames.
Similar Approaches
The idea that an event or state can stand metonymically for another event or
state within the same frame, or frame-like mental structure, and that one of the
events or states can be generalized, or schematized, to a concept that exists
independently of the original frame (thus leading to a metaphor) is compatible
with several approaches that provide accounts of how metaphors and meta-
phorical meanings emerge.
One such approach is Grady and Johnson’s (2002) work, where they note that
some metaphorical usages are associated with scenes where the literal and
metaphorical meanings of words (like the visual and cognitive senses of “see”)
are not initially distinguished by children; the two meanings of a word are
conflated. Later on, however, the meanings are distinguished by children, that
is, deconflated. The notions of conflation and deconflation are Grady and
Johnson’s way of accounting for the emergence of primary metaphors.
A primary scene, such as that of “becoming aware by seeing,” consists of two
subscenes: in this case, the perceptual subscene of seeing something and the
cognitive subscene of becoming aware of something. For example, in the process
of the emergence of the metaphorical sense of “see,” the child learns to differen-
tiate the cognitive subscene and the perceptual one, which gives rise to the
primary metaphor knowing/ understanding is seeing. In the account
I am proposing, the stage of deconflation could be seen as being preceded by a
metonymic stage, where one of the events or states stands metonymically for the
other within the same primary scene. Furthermore, I would also suggest that
deconflation can occur after the generalization, or schematization, of one of the
events or states has taken place. This stage can be seen as opening the way to the
emergence of metaphors – the stage where an event or state is generalized to a
44 Direct or Indirect Emergence?
(1) This means that while Julia is cooking she can still keep an eye on their
two young children.
(2) [The job] entails collecting the rent and keeping an eye on some housing
association flats.
(3) Often a scheme will need backing for several years. During this time the
Field Director [will] keep an eye on its progress.
The idiom keep an eye on has to do with the conceptual metaphor control-
ling is seeing. Deignan argues that in the first example keep an eye on is
metonymic, in the second it is ambiguous between a metaphoric and meto-
nymic interpretation, and in the third it is metaphoric. Example (2) that is
ambiguous between a metaphor and a metonymy corresponds to the deconfla-
tion stage and example (1) to the metonymic stage in my interpretation. It
should be noted, however, that this corpus-linguistic evidence is only partial. It
relates to the contemporary simultaneous existence of the metonymic,
metonymic-metaphoric, or metaphoric uses of the expression. It does not tell
us anything about the sequence in which this differentiation may have evolved
through time.
Several other authors make a more or less similar argument concerning the
metaphor-metonymy relationship, but they all use conceptual machinery that is
different from the one I use here (see, e.g., Barcelona 2000; Taylor 1989; and
several papers in Dirven and Pörings 2002). For example, a piece of work that
seems fully compatible with the approach advocated here is Radden’s (2002)
paper. Radden argues that in a large number of cases metaphors derive from
metonymies and that they emerge through a continuum involving literal
meaning through metonymy to metaphor. He distinguishes four cases where
metaphorical meaning, and more generally, metaphors, can derive from meton-
ymy: (1) correlations in experience, (2) conversational implicature, (3) taxo-
nomic structure of categories, and (4) cultural models. The obvious similarity
between his approach and mine is that we both see many metaphors as
deriving from metonymy. As a matter of fact, several of the examples dis-
cussed in his paper are the same as the ones discussed here. An important
difference, though, between his proposal and mine is that while Radden
3.3 Some Further Issues 45
identifies four different sources for the emergence of metaphors from meton-
ymy, I offer a more unified view that crosscuts his four sources and that
accounts for metonymy-based metaphors in terms of two events or states being
correlated in experience, a situation that is represented conceptually by means
of a frame-like mental structure. In other words, I view correlations between
two events or states as involving or underlying three of his four proposed
metonymic sources. (Radden also recognizes correlation as a metaphor-
producing source in his first group of sources.)
The view of the metonymy-metaphor relationship as based on correlations
between elements of a frame would bring the account I propose closer to
Grady’s notion of primary metaphor that is also based on the notion of
correlation. However, Grady’s view and mine are different on a number of
points. Grady and Johnson (2002: 540) define primary metaphors as follows:
“. . . primary metaphors are motivated by tight correlations between distin-
guishable dimensions of recurring, locally defined experience types. We refer
to these dimensions, which unfold dynamically over very brief time spans, as
subscenes.”
And in the same paragraph they continue: “Note that because subscenes are
co-occurring aspects of simple scenarios, our account might be taken to
suggest that primary metaphors arise from metonymies; it is important to
consider, however, that a metonymic relationship concerns conceptual and
referential association, whereas our proposal refers to correlations at the level
of experience, and to truly metaphoric patterns of conceptualisation which
arise from these correlations.”
Here Grady and Johnson explicitly deny the possibility that primary meta-
phors arise from metonymies – the claim I have made. They base their
argument on the idea that while metonymy “concerns conceptual and referen-
tial association,” primary metaphors concern correlations in experience. But in
order to become linguistically coded, experiential correlations must be con-
ceptualized, and once conceptualized, experiential correlations also become
conceptual (associations). Grady and Johnson also argue that primary scenes
and the subscenes that constitute them have a unique status among the
conceptual tools (such as domains, frames, mental spaces, schemas) used to
represent categories in the conceptual system. If valid, this argument could be
used to suggest that since metonymies can be found in frames, and since
primary scenes are different from frames (and other similar conceptual
structures), primary metaphors cannot derive from metonymies (which are tied
to frames). However, it is notoriously difficult to distinguish frames, mental
spaces, domains, schemas, scenarios, scripts, and, importantly, scenes from
each other in such a way that a distinction between primary metaphor and
metonymy could be based on the distinction. All of these conceptual structures
are mental representations of coherent aspects of human experience. As such,
46 Direct or Indirect Emergence?
despite their various differences in abstractness and the kinds of aspects of the
world they are taken to represent in the various disciplines (see, e.g., Andor
1985), they are all coherent mental representations of aspects of the world
consisting of parts and being conceptualized as wholes. (This is why in several
places in the chapter I have used the term “frame-like mental structure” that fits
the general, overarching definition. In the next chapter, Chapter 4, I will take
up this issue for a different purpose.) The correlations between events, states,
actions, etc. (the subscenes) that make up primary scenes may be thought of as
coherent organizations of human experience (i.e., frame-like conceptual struc-
tures) as well, with parts constituting wholes. Because of this, metonymies
may be based on primary scenes as well, and, as proposed above, may in turn
lead to metaphors (for an early formulation of this idea, see, e.g., Kövecses,
2002/2010; Kövecses and Radden 1998; Radden and Kövecses 1999).
In a more recent paper, Grady (2005) is careful to point out that many
metonymies are also correlations between concepts (such as author for
work, place for institution, means for action), or to put it differ-
ently, some correlations lead not to metaphoric connections but to metonymic
ones. Grady proposes that those correlations in experience lead to primary
metaphors that meet three conditions: (1) the correlation is between a sensory
and a nonsensory (mental) experience; (2) the two correlated experiences share
highly schematic structure; and (3) the correlated experiences covary, that is, if
there is a change in one, there is a similar change in the other.
In other words, on this view, a small group of correlations (characterized
by the three properties) gives rise to metaphor, not to metonymy. As an
alternative account, I would propose to consider this special correlation type
as deriving from metonymy because, as I argued, the components of primary
scenes function as parts (elements) within a single frame-like mental struc-
ture conceptualized as a whole, where a component part can provide access
to either another component part or to a whole (and then, through
generalization, paving the way for the emergence of full-blown cases of
metaphor). As a matter of fact, the sensory-nonsensory condition that char-
acterizes metaphors in the theory of primary metaphor may well be thought
of an equally good condition for metonymy. Given a frame or (primary)
scene, the conceptual element that represents something sensory in the frame
or scene is typically selected to stand for something nonsensory. That is, the
sensory-nonsensory distinction can be regarded as a further member of the
series of factors (or conditions), such as “human over nonhuman,” “visible
over nonvisible,” “specific over nonspecific,” etc., that are commonly evoked
to explain which element can provide access to another element in the most
typical manifestations of metonymy (Kövecses and Radden 1998; Radden
and Kövecses 1999).
3.3 Some Further Issues 47
Components of Scenes
I suggested above that scenes are like other frame-like mental structures that
can lead to metonymy (and then metaphor). The question I would like to take
up now is what the components are, or can be, that make up such frame-like
structures. One way in which primary scenes are distinguished from other
frame-like structures is that they consist of two subscenes of events, actions, or
states, both involving at least one entity and one relation (e.g., x sees y and x
knows y). These events, states, and the like (the subscenes), are then taken to
be correlated and form the basis of primary metaphors. But can we distinguish
primary scenes from other frame-like representations on this basis? Is it the
case that frame-like mental structures that can give rise to metonymy consist of
entities (representing individuals and relations) only and that they cannot
include (pairs of ) combinations of entities and relations (i.e., subscenes) of
the kind we have seen, for example, for knowing is seeing? This is not the
case. Recent metonymy researchers commonly work with notions like com-
plex events and subdomains in their studies of metonymy, which often contain
combinations of entities and relations (similar to “x sees y”). In other words, it
appears that the proposed nature of subscenes does not distinguish primary
scenes from other frame-like mental structures. It would perhaps be more
expedient to think of frame-like representations as being composed of either
individual entities and relations, combinations of entities and relations, com-
plex events consisting of simple ones (such as cooking or speaking), or
subdomains constituting larger domains. For the purposes of the study of
metonymy, we could conceive all of them as elements of frame-like structures
that can give rise to metonymies. The elements can be of any complexity and
either one can be used for another element or the whole (given the conditions,
including “sensory over nonsensory,” mentioned in the previous subsection).
However, it is clear that for metaphor to emerge from such frame-like struc-
tures we need two elements that have more complexity than simple entities, as
Grady and Johnson correctly observe. But this complexity may derive not only
from subscenes but also from any complex events or subdomains.
correlations in experience can also bring together elements (of any complexity)
in frame-like mental structures and can, at least initially, produce metonymies
within the frame, and, then, metaphors that are based on the metonymies. It
might turn out that some of our concepts that form conceptual metaphors today
may be (or may have been) based on correlations in experience that are less
tight and that have worked through historical time (e.g., correlations between
symbols and their referents, between models and what they are models of ).
The evolution of particular cultures and systems of categories can probably
also produce metaphors that emerge over long time spans through a meto-
nymic stage in the process. If this suggestion is valid, a number of conceptual
metaphors that are now taken to be nonprimary and not based on correlations
may turn out to be correlational ones, thus further blurring the distinction
between primary and nonprimary metaphors.
3.4 Conclusions
In the chapter, I suggested that many metaphors arise from, and are not
independent of, metonymies. It should be made clear that, contrary to some
interpretations, the claim is not that correlation metaphors are not really
metaphors but metonymies. For example, Dancygier and Sweetser (2014:
104) write in connection with my suggestion: “Kövecses (2013) explicitly
argues that all conceptual metaphors are based in metonymy; however, correl-
ation is not metonymy but the basis for metonymic uses of one thing to stand
for something correlated with it.” This is not what I claim. First, I do not claim
that all conceptual metaphors are based in metonymy. Obviously, resemblance
metaphors are not metonymy-based. Second, my proposal was that correlation-
based metaphors emerge from frame-like mental representations through a
metonymic stage (not a metonymic linguistic expression), a “metonymy-pro-
ducing relationship” (see Kövecses and Radden 1998; Radden and Kövecses
1999) between two elements of a frame (if you wish, “the basis for metonymic
uses”). This is, I assume, a perfectly natural process. It is in line with how
correlations in general produce a large number of metonymy types. Indeed, we
can think of experience types as being constituted by several more basic
correlated elements (parts). Given the elements of experience constituting
experience types, there are a large number of metonymy-producing relations
that underlie the metonymies (metonymic expressions) we use.
When do the elements that are metonymically related lead to metaphors?
I proposed this happens when one of the elements of a frame-like mental
structure is generalized (schematized) to a concept that lies outside the initial
frame in a different part of the conceptual system. The generalization process
leads to sufficient conceptual distance between the initial and the new frame on
which metaphors can be based. In order for metaphors to emerge, in the typical
3.4 Conclusions 49
cases, distant vertical taxonomies are involved. This explains the importance
of taxonomic hierarchies and frames in our conceptualization of the “meta-
phor-metonymy relationship.”
The generalization (schematization) process can apply to either the element
in a frame that becomes the source or the target domain of a metaphor. When it
applies to an element that becomes the source (Case 1), the other element will
be the target of the metaphor (e.g., sadness is down). I described this
situation as “target giving rise to the source.” In the other case (Case 2), the
generalization applies to an element that becomes the target, and the other
element will be the source (e.g., knowing is seeing). I described the
situation as “source giving rise to the target.” Furthermore, when either the
source or the target is generalized (schematized), we can get elaborations. For
example, heat as the source of anger can be elaborated as “fire,” “stew,”
“boiling water,” “blood,” “piss,” “volcano,” “lava,” and so on. And when, say,
difficulty as target is elaborated in the metaphor difficulty is weight,
we get “problems,” “emotions,” “responsibilities,” and so on (and hence we
get the metaphors problems are weights, emotions are weights, and
responsibilities are weights). Needless to say, the source weight can
also be elaborated (as, for instance, “burden” and “baggage”) – in, often,
culture-specific ways.
I believe the view outlined here provides a coherent overall account of why
many metaphors can be interpreted as metonymies and why many metonymies
can be taken to be metaphors. The reason seems to be that, in the case of
correlational metaphors, between a nonfigurative and a metaphoric stage there
is a metonymic one. Given a frame in which the correlations apply between
two elements, there must be a logically prior metonymic relationship between
the elements before one element is moved out of the initial frame to establish a
new and independent frame, which will form either a metaphoric source or
target. Only future research can decide whether this explanation can be
extended to cases of conceptual metaphors not analyzed (only hinted at) in
the present study.
4 Domains, Schemas, Frames, or Spaces?
1
In the first half of the chapter, I reuse materials from Kövecses (2017b).
50
4.1 Image Schemas, Domains, Frames, and Spaces 51
a theoretical construct in CMT? Finally, one could select just a pair of these
terms (such as image schema and domain or mental space and domain) and
claim that these are sufficient to describe and explain conceptual metaphors. It
seems then that there is a very real need to raise the issue of which conceptual
structure or structures (or unit(s)) constitutes (or constitute) conceptual meta-
phors. Given this situation, my main goal in the chapter is to answer the
following general question:
What is (are) the appropriate conceptual structure/unit (or structures/units)
involved in conceptual metaphors?
My suggestion will be that it is best to think of conceptual metaphors as
simultaneously involving conceptual structures, or units, on several distinct
levels of schematicity (see, e.g., Lakoff 1987; Langacker 1987; Rosch 1978).
I distinguish four such levels: the level of image schemas, the level of domains,
the level of frames, and the level of mental spaces (in addition to the linguistic
level of the actual utterances in which the metaphors are instantiated). My
claim is that the resulting overall picture of conceptual metaphors provides us
with a new comprehensive framework for the study of metaphor in conceptual
metaphor theory. I would like to refer to the emerging new framework as the
“multilevel view of conceptual metaphor.” To demonstrate my points, I will
mainly use the source domain of building in this chapter (see, e.g., Grady
1997a,b; Kövecses 1995, 2000a,b, 2002/2010; Lakoff and Johnson 1980).
In what follows, first, I describe the theoretical background, where I justify
the relatedness of the four levels mentioned above, independently of metaphor.
Second, based on these findings, I apply the results to the study of conceptual
metaphors in general. Third, I look at the well-studied source domain of
building in some detail in order to show how the levels work together in
this particular example. Fourth, I summarize the main points and describe the
general picture of the multilevel view of metaphor. Fifth, I examine some of
the consequences of the new view for a variety of issues related to metaphor.
Finally, I offer some conclusions, together with indicating some potential
benefits and future directions of research involved in the new model.
IS
DM
FR
MS
In the diagram, IS stands for image schema, DM for domain matrix, FR for
frame, and MS for mental space.
Image schemas are essential conceptual structures that imbue experience
with meaning (cf. Johnson 1987; Lakoff 1987). Based on the literature on
image schemas, Hampe (2005: 1–2) finds four features of image schemas
especially characteristic. I present them here in a somewhat simplified form.
Image schemas are
directly meaningful preconceptual structures;
highly schematic gestalts;
continuous analogue patterns;
internally structured, consisting of only a few parts.
Because of their highly schematic nature, image schemas range over the entire
conceptual system making a wide variety of concepts and experiences mean-
ingful. For example, the concept of journey presupposes the more schematic
structure of motion and, more specifically, source-path-goal motion
(to distinguish it from other types of motion). Further, concepts may take
several image schemas to support them conceptually. For example, the concept
of body is based on the image schemas of container, verticality,
(structured) object, and so on. Finally, concepts may also be character-
ized by the same image schemas. For instance, the concept of building (in
the sense of an enclosed construction), similar to the body, presupposes the
container, verticality, and object schemas.
The notion of domain is defined by Langacker (1987: 488) as follows:
“Domain [is] A coherent area of conceptualization relative to which semantic
units may be characterized.” Unlike image schemas, domains are not analogue,
imagistic patterns of experience but propositional in nature in a highly sche-
matic fashion. They are at a level immediately below image schemas. Image
schemas make domains such as journey, body, and building (mentioned
above as concepts) meaningful. The various image schemas apply to different
aspects or dimensions of a domain in the sense of a domain matrix (Langacker
1987). A domain as a domain matrix (such as building) presupposes a
variety of concepts that characterize different aspects of the domain. Domains
have many more parts than image schemas, and are thus more information-
rich. The definition of domains does not distinguish between domains and
frames, or idealized cognitive models (Lakoff 1987), as noted by Langacker
himself. The only way to distinguish the two is in terms of schematicity.
Frames (see, e.g., Fillmore 1982) are less schematic conceptual structures
than domains. As noted above, the definition of domain by Langacker can
serve as a definition of frame as well. (On the blurred boundaries between
domains and frames, see also Cienki 2007.) The difference between a domain
and a frame, in my view, can be captured by a difference in schematicity
54 Domains, Schemas, Frames, or Spaces?
between the two: Frames elaborate particular aspects of a domain matrix; that
is, particular higher-level concepts within a domain. Sullivan (2013) considers
this relationship as inclusion. According to her, domains include or consist of
frames. We can also think of the relationship between the two as different
degrees of schematicity. It can be suggested that frames involve more concep-
tually specific information than domains. For example, the body domain can
be seen as being elaborated by several distinct frames, such as perception,
ingestion, and exercising (see Sullivan 2013). These frames account for
such metaphorical linguistic expressions as I see what you mean (percep-
tion), digest an idea (ingestion), and a mental exercise (exercising)
(Sullivan 2013). Together, they make up what is known as the generic-level
metaphor the mind is the body (see Johnson 1987; Sweetser 1990). In
general, the frames elaborating a domain consist of roles and relations between
the roles and the roles can be filled by particular values.
When the roles are filled by particular values in actual discourse in specific
communicative situations, we have to do with mental spaces (see Fauconnier
1994). I use mental spaces in the sense the term is defined by Fauconnier
(2007: 351): “Mental spaces are very partial assemblies constructed as we
think and talk, for purposes of local understanding and action. They contain
elements and are structured by frames and cognitive models. Mental spaces are
connected to long-term schematic knowledge, such as the frame for walking
along a path, and to long-term specific knowledge, . . .” This is also what
Langacker (2008) calls “current discourse space” (CDS). In Langacker’s
words: “The CDS is a mental space comprising everything presumed to be
shared by the speaker and hearer as the basis for discourse at a given moment”
(Langacker 2008: 59). Mental spaces borrow their structure from frames, but
the generic structures from frames are further elaborated by specific infor-
mation from context. In addition, mental spaces can be structured by one or
several different frames. That is, they can be the realizations of a single frame
or they can rely on a combination of roles and relations from several distinct
frames. Mental spaces are, then, even more specific than frames, in that they do
not operate with generic roles and relations in most cases but with specific
instances of roles and relations. At the same time, they are also coherent
organizations of experience, just like frames and domains, but they function
at a very specific and conceptually rich level. Moreover, mental spaces are
used in online processing for purposes of local understanding, as noted above,
and also by Fauconnier and Turner (2002), but, as we just saw, they are
structured by frames. Mental spaces are online representations of our under-
standing of experience in working memory, whereas frames and domains are
conventionalized knowledge structures in long-term memory.
In brief, image schemas, domains, frames, and mental spaces are all used
by conceptualizers/speakers for the purposes of lending organization and
4.1 Image Schemas, Domains, Frames, and Spaces 55
SH 1 SH 2 SH 3 SH 4 Schematicity hierarchies
L1 Image schemas
L2 Domain matrixes
L3 Frames
L4 Mental spaces
Figure 4.3 The schematicity hierarchy with three major distinctions between
the levels
say, building, as we will see in the next section, represents our entire
knowledge structure associated with this concept. In general, out of the entire
domain matrix, only select aspects of the domain matrix participate in the
mappings between source and target concepts.
The aspects that do participate in the mappings can be, and usually are,
given in the form of frames – at a lower level of schematicity. The frames
elaborate the select aspects of domains. Thus, the source frames offer more
specific information than domains, but they do not cover or exhaust all aspects
of a source domain (matrix).
As regards metaphorical mental spaces, they are the least schematic con-
ceptual structures of the four discussed here (i.e., image schemas, domains,
frames, and mental spaces). They are highly specific structures occurring in
online processing in particular communicative situations. As such, they are
fully influenced by several different kinds of contextual information (Kövecses
2015a). Mental spaces participating in metaphor can be seen as elaborations of
frames that elaborate select aspects of domains for metaphorical conceptual-
ization. They contain the most specific information that derives from filling out
generic roles with particular values, as well as from the specific context in
which the spaces emerge.
While domains and frames belong to the level of cognitive organization and
analysis of metaphor that Kövecses (2002/2010) calls the “supraindividual”
level, mental spaces occupy a level that can be thought of as the “individual”
level. In a previous publication (Kövecses 2002/2010), I distinguish the
supraindividual, individual, and subindividual levels in the following way:
“In conclusion, then, the cognitive linguistic view of metaphor that has been
discussed in this book works on three levels: the supraindividual level corres-
ponding to how a given language and culture reflects decontextualized meta-
phorical patterns, the individual level corresponding to the metaphorical
cognitive system as used by individual speakers of a language, and the
subindividual level corresponding to universal aspects of various kinds of
embodiment.” (Kövecses 2010: 321)
The supraindividual level corresponds to the level of domains and frames. In
the CMT literature, descriptions of conceptual metaphors are typically given at
the supraindividual level. Such descriptions lack the specificity and richness of
metaphorical conceptualization that we find at the individual level – the level
corresponding to the use of mental spaces in the present framework.
The idea that mental spaces can be thought of as the most specific concep-
tual structures participating in metaphorical conceptualization finds some
support in more recent work. Musolff (2006) uses the term “scenario” in the
same sense in which I use “mental space.” (Actually, Musolff also roughly
equates his use of scenario and scene with mental space.) In discussing a
particular example, he writes:
58 Domains, Schemas, Frames, or Spaces?
In other words, Musolff also suggests that scenes and scenarios function below
the level of domains and frames, where speaker and hearer metaphorically
conceptualize their experiences in a fully contextualized fashion. The speaker
and hearer take advantage of a large amount of knowledge available to them in
the specific communicative situation. His observations are thus very much in
line with the present approach that emphasizes the various degrees of sche-
maticity in metaphorical conceptualization.
We can represent the emerging picture of metaphorical conceptualization in
a schematic way in the diagram to follow:
Source Target
individual who built the house and a structural property of the house. These
are pieces of information that are not contained in the more schematic
building frame.
Now let us look at the commonly cited metaphorical examples that can be
found in the CMT literature in connection with the building source domain.
I will mainly use here what I call the “lexical approach” to collecting metaphors
that are characteristic of a particular domain such as building (see Kövecses
1986, 2015b). In essence, the lexical approach takes advantage of a variety of
lexical materials (idioms, collocations, synonyms, etc., as these are available in
dictionaries and other collections of lexical items) in finding metaphorical
expressions. The method performs well as regards collecting conventionalized
and decontextualized types of metaphorical expressions; it performs poorly as
regards finding novel types and tokens of metaphors and quantifying the
metaphorical data in relation to a domain. Since in the chapter I am concerned
with the conventional types of metaphorical expressions (and not with quanti-
fying metaphor occurrence in actual use), I employ the lexical method (see also
the subsection on methodology discussed later in this chapter). Nevertheless,
I will also rely on some corpus examples in the analysis to follow.
We can begin with Lakoff and Johnson’s Metaphors We Live By, where we
find this for the metaphor theories are buildings (Lakoff and Johnson
1980: 46):
theories (and arguments) are buildings
Is that the foundation for your theory?
The theory needs more support.
The argument is shaky.
We need some more facts or the argument will fall apart.
We need to construct a strong argument for that.
I haven’t figured out yet what the form of the argument will be.
Here are some more facts to shore up the theory.
We need to buttress the theory with solid arguments.
The theory will stand or fall on the strength of that argument.
The argument collapsed.
They exploded his latest theory.
We will show that theory to be without foundation.
So far we have put together only the framework of the theory.
For the related concept of argument, we find the following examples:
an argument is a building
We’ve got the framework for a solid argument.
If you don’t support your argument with solid facts, the whole thing
will collapse.
4.3 The Example of building 61
Essentially, it is these meaning foci that, with the exception of the first,
Grady’s (1997a,b) primary metaphors also capture: logical organiza-
tion is physical structure (corresponding to 2) and persistence is
remaining erect (corresponding to 3). In other words, if this analysis is
correct, we can suggest that primary metaphors and the main meaning focus
are, in part, related phenomena: When the mappings of a conceptual metaphor
are based on primary metaphors such as the ones above, these will determine
the main meaning focus of the complex metaphor (like theories are
buildings). (However, it should be noted that the notion of main meaning
focus also applies to non-primary-metaphor-based, i.e., analogy-based, con-
ceptual metaphors, such as computers are people or people are com-
puters, where people in the former metaphor are associated with, e.g.,
emotionality, while computers in the latter with, e.g., a high level of memory
capacity and mechanical calculations.)
In light of the analysis so far, it seems that the building source domain
participates in the conceptualization of theories (and, more generally, of
complex abstract systems) by means of two frames within it: the
building (as process) frame associated with the construction aspect
and the (physical) support frame with the structure aspect of the
building domain. The building (as process) frame gives rise to the first
mapping above, “building ➔ creation.” The frame utilizes the conceptual
elements of the builder, the action or activity of building, and the thing that
gets built. Thus, the builder corresponds to the person who creates or
constructs the theory, the building action to the creation or construction
process that results in the theory, and the thing built to the theory itself.
Interestingly, it does not conventionally utilize the materials that are used
in the building activity – at least according to the examples given (but
see below).
The second frame, (physical) support, is represented by the second
and third mappings above. In these mappings, we have two important
structural elements that are related to each other in a particular way, as well
as a property that is expected to characterize both the elements and the
structural relationship between them. The structural elements are the founda-
tion and the outer shell (of the building), and the relationship that obtains
between them is that the outer shell is supported by the foundation. The
property that characterizes the elements and the relationship between the two is
that of strength. These are the conceptual elements that are utilized in the
(physical) support frame.
The emphasis on this latter aspect of buildings has a long cultural history.
It is this relationship (supported by) and this property (strength) that are
emphasized, for example, by one of the parables in the New Testament:
4.3 The Example of building 63
(1) 25 The rain came down, the streams rose, and the winds blew and beat against that
house; yet it did not fall, because it had its foundation on the rock. 26 But everyone who
hears these words of mine and does not put them into practice is like a foolish man who
built his house on sand. 27 The rain came down, the streams rose, and the winds blew
and beat against that house, and it fell with a great crash.” (Matthew 7:25–27 New
International Version (NIV))
This may be a culturally significant factor that may have created, or, simply,
may reinforce, the conceptual use of these two specific frames within the
building source domain. The other two frames mentioned above, parts
of the building and function of the building, do not appear to
produce conventionalized metaphorical language on a par with building (as
process) and (physical) support, as evidenced by the examples used in the
literature and the dictionaries consulted.
The meanings of the conventionalized expressions, the three mappings on
which they are based, the two frames associated with the building source
domain, as well as the other constitutive frames are at, what I called, the
supraindividual level (i.e., in long-term semantic memory), that is, at the level
where we store decontextualized conceptual information related to building
as a conceptual domain.
The issue we consider next is what happens when this body of information
is actually used in real discourse at the individual level, that is, where actual
people communicate in real contexts characterized by a great deal of specific
information available to them.
preparation and planning, the application of the building frame can also be
extended to such aspects of the building process. I assume that this is what
Lakoff and Johnson (1980/2003) refer to as the “extension of the used part” of
the metaphor. They exemplify it with the sentence “These facts are the bricks
and mortar of my theory,” where bricks and mortar constitute the frame
element of building materials. Additional examples are not difficult to find.
A simple Google search returned the following example: “This middle level
between theories and concepts allows models to serve a critical function within
science; they act as the bricks and mortar of a theory and are the basis for how
scientists argue” (www.ncbi.nlm.nih.gov/pmc/articles/PMC3671648/). The
use of the building frame with the extension to building materials may not
be frequent and conventional, but when speakers conceptualize theories meta-
phorically in actual discourse they may be prompted by the frame to extend its
application to this frame element (by way of completing the frame).
Lakoff and Johnson (1980/2003) also mention another example: “His theory
has thousands of little rooms and long, winding corridors.” They refer to it as a
case in which the metaphor derives from an “unused part” of the source
domain. We can now provide a fuller account of such cases within the
framework I am developing here. It can be suggested that in cases like this
frame elements from other than the routinely or conventionally used frames
can participate in metaphorical mappings. The cognitive mechanism that
seems to be involved can be characterized as follows. Again, Musolff (2001)
provides several pertinent examples, which I reanalyze here (based on
Kövecses 2005):
“We are delighted that Germany’s unification takes place under the European roof.”
(Documentation by the Federal press- and information office, Bonn)
“At the moment, the German occupants of the first floor apartment in the ‘European
house’ seem to think that foreigners from outside the continent should be content with
living in the rubbish bin.”
“What does he [Chancellor Kohl] need this house for, after so many years as
Chancellor? – Well, it’s obvious, he wants to become the caretaker.” (Die Zeit, May
16, 1997)
“[the European house is] a building without fire-escapes: no escape if it goes wrong”
(The Guardian, May 2, 1998)
“[it is a] burning building with no exits” (The Times, May 20, 1998)
Musolff notes that, given the common european house metaphor intro-
duced in public discourse in the 1990s, i.e., the european union is a
building metaphor, all kinds of metaphorical expressions showed up in
political discourse at the time that were different from the ones presented at
the beginning of the section (i.e., were outside the profile, or meaning focus, of
the source). In addition to the highly conventionalized linguistic metaphors,
such as lay the foundations, building, buttress, framework, and so on,
4.3 The Example of building 65
CAS
FR PHYSICAL
BUILDING SUPPORT CONSTRUCTION LOGICAL SUPPORT
architect foundation designer framework/organization
PARTS OF plans FUNCTION OF supports designs is based on/supported by
BUILDING builder BUILDING shell builder basic ideas
builds strong builds supported in a stable way
building organization/framework
? from materials ? from ideas
The diagram above shows both the schematicity hierarchy of the source and
target concepts (from IS to MS) in a vertical arrangement and the mappings
between the source and target concepts on the four levels in a horizontal
arrangement. The concepts constituting the conceptual metaphors on the
various levels range from most schematic to least schematic. Thus we get the
following conceptual metaphors:
IS level:
complex abstract systems are objects
complex abstract systems are vertical objects
complex abstract systems are container objects
complex abstract systems are whole objects with parts
DM level:
a (complex abstract) system is a physical object (building)
the creation of the system is the physical creation of the
building
the structure of the system is the physical structure of the
building
the parts of the system are parts of the physical object (building)
FR level:
the construction of the system is building (as process)
logical support is physical support
submetaphor:
the stability of the building is the strength of physical support
MS level:
(in the form of particular examples:)
mary building a career is mary building a house
john’s life without a foundation is john’s house without a
foundation
These mappings at the different levels license the use of metaphorical expres-
sions at different levels of schematicity. For example, the word inside, as in
“inside the system,” relates to the image schematic level (the container
image schema), while the word structure (when used metaphorically) is based
on a mapping at the domain level. Similarly, the word builder in the phrase
“the chief builder of the organization” is related to the building (as process)
frame. In other words, metaphorically used words, just like their corresponding
conceptual metaphors, can be found at various levels of schematicity.
only coherent with their targets (or the other way around), but also with context
in a broad sense. Elsewhere, I proposed that many metaphorical source
domains are coherent with one or more contextual factors. This constraint
was called the “pressure of coherence” with the context (see Kövecses 2005,
2015a). There seems to be abundant evidence for this kind of activation in the
experiments conducted by Boroditsky, Casasanto, Gibbs, Colston, and other
researchers (see, e.g., Boroditsky 2001; Casasanto 2009; Gibbs 2006; Gibbs
and Colston 2012.) In their study of figurative language, Dancygier and
Sweetser (2014) make a proposal that is close to the present one as regards
the emphasis they place on different levels of schematicity in accounting for
metaphoric language. David, Lakoff, and Stickle’s (2016) notion of “cascades”
is a formalized way to capture the same phenomena. What distinguishes the
present approach from these is that (1) in addition to image schema, domain,
and frame, I include the notion of mental spaces in the hierarchy and (2)
I consider context to play an essential role in the emergence of the hierarchies.
(I discuss the role of context in Chapter 5.)
All in all, then, we have the following picture:
analogue
structure
IMAGE SCHEMAS
…………………………………………………………………………
long-term memory
DOMAINS
non-analogue structure
FRAMES
------------------------------------------------------------------------------------
MENTAL SPACES
working memory
George Lakoff, (2) a distinction emphasized by Gerard Steen, (3) the general
issue of metaphorical idioms, and (4) a fundamental issue in the methodology
of studying metaphor within a CMT framework. In particular, they are as
follows:
What are “deep” versus “shallow” or “superficial metaphors? What is their
ontological status in the metaphorical conceptual system?
Do “deliberate” metaphors exist independently of “nondeliberate” ones?
How best can we characterize them in a multilevel view of metaphor?
How does the multilevel view of metaphor apply to metaphorical idioms?
Does it shed any new light on their nature?
If visual metaphors are the products of our metaphorical conceptual systems,
which I assume to be hierarchically structured, visual metaphors must also
be hierarchically structures. Is this indeed the case?
What is the “best” method to study metaphors out of the many available
ones in the various disciplines that deal with metaphor?
In what follows, I address the four questions and attempt to examine them in
light of the multilevel view of metaphor.
become a “hammock” that is robbing the country of freedom and virtue” (from
Lakoff 1995).
Lakoff offers an interpretation of the metaphor along the lines of his book
Moral Politics (1996). He states that “[t]he tightrope is straight and narrow – a
moral path.” This is based on the well-known conceptual metaphor: moral is
straight. Given the metaphor, walking the tightrope corresponds to working
and falling off to losing your job.
In addition to this interpretation, several other metaphors seem to be equally
possible in understanding the passage. One of these would be the conceptual-
ization of actions in general as motion. Another would take into account the
balancing act of the tightrope walker. Furthermore, we can think of the
movement of the tightrope walker as a life-related metaphor having journey
as its source domain. Finally, we can interpret the situation depicted by the
sentence as a combination of control and life-action metaphors, in which
controlling life amounts to keeping a job. The set of conceptual metaphors
according to this interpretation would thus be:
action is motion
control is physical balance
life is travel
leading a life is journeying
keeping one’s job is keeping one’s balance while walking
along a tightrope
losing one’s job is losing one’s balance while walking along
a tightrope
The last two metaphors are the most specific ones, given the schematicity
hierarchy mentioned above. They occupy the level of mental spaces. These
highly specific metaphors go together with some everyday knowledge that is
carried over from the source to the target:
The tightrope walker needs a safety net so as not to injure himself if
he falls.
People need social support in case they lose their jobs.
The knowledge provides the basis for the social support of working people.
More importantly for the argument here, the conceptual metaphors identi-
fied above (i.e., control is balance, action is motion, etc.) can be
found at four different levels: the levels of image schema, domain, frame, and
mental space in the following way:
Image schema:
action is motion
control is physical balance
Domain:
life is travel
4.5 Some Consequences of the Multilevel View 73
Frame:
leading a life is journeying
Mental space:
keeping one’s job is keeping one’s balance while walking
along a tightrope
losing one’s job is losing one’s balance while walking along
a tightrope
The schematicity hierarchy works in the following way in this case: Actions in
general are metaphorically understood as (self-propelled) motion (image
schema level). Life is a kind of action and travel is a kind of motion (domain
level). It is a major aspect of life that we live (lead) it, which is understood
through journeying (the activity aspect of travel; see Framenet) (frame level).
A part of leading a life is having (or not having) a job, which is understood as
walking along a tightrope (mental spaces level). The motion and balance
source domains are conceptually integrated (blended) at the level of mental
spaces in the image of “walking the tightrope” (on blending, see Fauconnier
and Turner 2002). As can be seen, both the target and source domains are
more specific elaborations of higher-level metaphors, as we go down the
hierarchy.
This organization of the metaphors tells us about what the distinction
between “deep” and “superficial” metaphors actually means. The deep meta-
phors would be the image schema-, domain-, and frame-level ones. The
superficial metaphor would be the one at the level of mental spaces. In other
words, we can reinterpret and reconceptualize Lakoff’s distinction between
deep and superficial metaphors in terms of the schematicity hierarchy shown
above. The deep metaphors are the ontologically most schematic ones, while
the superficial ones are the least schematic. His terms “deep” and “superficial”
are metaphorical construals of the three “higher” (image schema, domain,
frame) and the “lowest” (mental space) levels.
having the lights on in a house, whereas its lack corresponds to darkness in the
house. In short, the metaphor at the mental-space level is: lack of normal
mental functioning/lack of understanding (i.e., having alz-
heimer’s disease) is not having the lights on in a house.
If this analysis is on the right track, we can suggest that deliberate metaphors
come with a large nondeliberate part at the image-schematic, domain, and
frame levels. The deliberate metaphor that is presented at the least schematic,
most experience-rich level (the level of mental spaces) is accompanied by and
actually assumes a set of systematic “higher-level” (in Lakoff’s sense,
“deeper”) conceptual metaphors and image schemas.
Another example that comes from my own work (Kövecses 2015a) and
points in the same direction can be offered. An article from USA TODAY is
about cyclist Lance Armstrong’s confessions concerning his doping, and tells
us that his confessions up to that point had not been sufficient to redeem
himself and clean up the sport of cycling. Several (crisis) experts who were
interviewed thought that additional steps must be taken by Armstrong to
achieve this. One specialist in crisis management said this in an interview:
“To use an analogy from the Tour de France, he’s still in the mountain stage,
and will be for some time” (2013, USA TODAY, 6W Sports, Weekly Inter-
national Edition, my italics, ZK).
What we have here is that the specialist has extensive knowledge about
the topic of the discourse, which is Armstrong’s doping scandal. That
knowledge includes that as a cyclist Armstrong participated in several Tour
de France events and that this race has several “mountain stages.” I suggested
that the topic of the discourse primed the speaker to choose a metaphor to
express a particular idea; namely, that, in order to come completely clean,
Armstrong has a long and difficult way to go. This idea was expressed by the
metaphorical linguistic expression in the mountain stage – a deliberate
metaphor. The metaphor is based on the mapping “impediment to motion
➔ difficulty of action (making full confession and being forgiven)” in the
action is self-propelled motion conceptual metaphor. But this is
insufficient as an account for the meaning of the metaphor because there is
still a wide conceptual gap between the meaning of the expression in the
mountain stage and the action is self-propelled motion conceptual
metaphor. The gap can be bridged if we assume the proposed multilevel
conceptual structures in which conceptual metaphors occur in a schematicity
hierarchy, such as the following:
Image schema:
Activities in general are metaphorically conceptualized as entities
moving. This yields the image-schema level metaphor:
76 Domains, Schemas, Frames, or Spaces?
ACTIVITY IS
This way of displaying the image-schema level metaphor is used to show the
analogue nature of metaphors at this level. The figure in source domain
position indicates an entity (square) moving (arrow).
Domain:
At the domain level we have the domains of communication, or
communicative action, and self-propelled (forward)
motion. They form the conceptual metaphor:
communicative action is self-propelled (forward) motion
Frame:
There are two frames participating in the conceptual metaphor at the
frame level: confession and race. The metaphor thus becomes:
confessions are races
Mental space:
At the mental space level, the conceptual metaphor is specified as a
bicycle race: confessions are bicycle races. What gets
focused on at this level is the length and difficulty of
confessing wrongdoings. Correspondingly, in the source
concept at this level we have length and difficulty of
being in the mountain stage of the tour de france.
This results in the conceptual metaphor:
the length and difficulty of confessing wrongdoings is the
length and difficulty of being in the mountain stage of
the tour de france
The metaphor is thus fully motivated – both conceptually (i.e., via being a part
of a schematicity hierarchy) and contextually (see Chapter 5). The two
examples I have briefly looked at above indicate not only that deliberate
metaphors have a large nondeliberate part at the image-schematic, domain,
and frame levels. They also show that the deliberate metaphors are actually
based on metaphors that are nondeliberate, and the deliberate and nondeliberate
parts are inseparable, they work jointly within the same schematicity hierarchy
of metaphorical concepts. Moreover, it follows that the metaphors that are
termed deliberate are really both deliberate and nondeliberate at the same time.
4.5 Some Consequences of the Multilevel View 77
One could argue, of course, that some deliberate metaphors are truly isolated
on the mental-space level; that is, unlike the previous examples, they do not
come with a large background conceptual package. Let us consider a potential
example for this possibility. We can take an example used by Steen, sonnet
18 by Shakespeare:
Shall I compare thee to a summer’s day?
Thou art more lovely and more temperate:
Rough winds do shake the darling buds of May,
And summer’s lease hath all too short a date;
Sometime too hot the eye of heaven shines,
And often is his gold complexion dimmed;
One could say that in the sonnet there is a single, isolated deliberate metaphor:
“Shall I compare thee to a summer’s day?” and that there are no deep metaphors
that would unconsciously guide the way we understand and appreciate this part
of the poem. However, a closer look reveals that the various layers of metaphors
that characterize the other examples can be found here as well.
Image schema:
intensity is heat / cold
Domain:
emotion is temperature: intensity of emotion is degree of
heat / cold
Frame:
love is fire: love’s intensity is the degree of heat of fire
Mental space:
the intensity of the poet’s love is the degree of a summer
day’s heat
Metaphorical Idioms
As a third illustration of how we can put to use the multilevel view of
metaphor, let us consider metaphorical idioms. It is well-known that particular
idioms can be underlain by conceptual metaphors (e.g., add fuel to the fire –
anger is fire). However, to think of metaphorical idioms in this way would
greatly oversimplify the picture. It seems that metaphorical idioms can be
assigned to very specific metaphorical ideas at one level, but this level is just
one layer of a complex hierarchy of schematicity. Moreover, the hierarchy can
end in a number of image-schema metaphors that jointly characterize the idiom
in question. Finally, for example, idioms, like idioms of money, can belong to
very different hierarchies with very different image-schema, domain-, and
frame-level metaphors (see Kövecses 2018, 2019).
Consider an example sentence taken from a dictionary: Everybody
borrowed money from Joe, the cash cow of the family. Here the metaphorical
expression is cash cow. Let us see what is involved in the metaphorical
conceptualization of the expression and the sentence in the multilevel view
of metaphor presented here.
At the most specific level, that of mental spaces, the person is identified as a
cow and the person’s money as milk. The money comes from the udder of the
cow, so that borrowing money amounts to getting milk from the udder. The
constitutive parts of the compound noun are cash and cow which are conceptu-
ally blended. Blending occurs at the mental spaces level. The dynamic mental
operation of conceptual integration can only take place at this level in working
memory. In sum, on this analysis, the expression cash cow assumes the
following very specific conceptual metaphors (corresponding to a set map-
pings between the source and the target) at the level of mental spaces:
joe is a cow
the money is milk
the source of the money is the udder of a cow
borrowing money from joe is getting milk from the udder of a cow
At the frame level, the last constituent metaphor becomes less specific: Joe
identified as a cash cow at the level of mental spaces is conceptualized as a
source-container and the milk as an object/substance (liquid). In addition, in
the prototypical borrowing-lending frame there is a borrower and a lender. The
borrower requests the money and the lender gives the money to the borrower.
Thus, we get:
borrowing and lending money is requesting the motion of an
object/substance (liquid) from a source(-container) and
the source(-container) causing the motion of the object/
substance (liquid) to the requester
4.5 Some Consequences of the Multilevel View 79
IS:
a person is a container
resources are substances
the source of resources is a container
occurrences are movements
change is motion
DM:
a person having money is a money-container
money is a liquid
the source of money is a liquid-container
change of possession is motion of substance out of source
FM:
the person who is requested to lend money is a source-
container
the money is a liquid-substance
borrowing and lending money is requesting the motion of
an object/substance (liquid) from a source(-container)
and the source(-container) causing the motion of the
object/substance (liquid) to the requester
MS:
the person is a cow
money is milk
the source of money is the udder of a cow
borrowing money from joe is getting milk from the
udder (and the cow giving it)
The description above of the metaphorical idiom cash cow in the given frame
(borrowing) shows the potential conceptual complexity of a single idiom.
Other idioms of money (e.g., cashflow) would require a different schemati-
city hierarchy due to the fact that they would involve other frames (such as
buying and selling). Still different hierarchies would be required by other
money-related idioms, such as “money talks,” where different image sche-
mas would apply (such as force). All these complexities could be fruitfully
studied by making use of the multilevel view of conceptual metaphors as
proposed in the book.
The notion of cubism, and more generally, art, as represented in the painting
by Tansey can be interpreted as a complex abstract system. Complex abstract
systems are routinely conceptualized metaphorically as complex physical
objects of various kinds. This yields the following highly schematic conceptual
metaphor for the cubism hierarchy: a complex abstract system is a
complex physical object.
At the image schema level, this metaphor can be represented in an analogue
fashion (i.e., here, visually) in the diagram below:
Vehicles are created, they have shapes (but not a uniform shape), they have
size and type, they are structured, they are made of some materials, they have
function, they have particular modes of operation, they work with various
degrees of force, and they can move. In short, the concept of vehicle comes
with an extensive domain matrix in Langacker’s (1987) sense.
At the frame level, we have cubist painting as target. The source concept
here emerges as an airplane – a special case of the concept of vehicle. The
4.6 Visual Metaphors 83
aspects of the source as defined by the domain matrix attaching to the concept
of vehicle become more specific here. For instance, airplanes are made in
certain ways, they have a characteristic shape, they are typically large, their
parts are assembled to form a unique structure, they are made of special
materials, and they can fly. These features apply to cubist painting and,
together, they result in the frame-level metaphor cubist painting is an
airplane.
What we know about these aspects individually and taken together greatly
contributes to the metaphorical understanding of cubism and cubist painting.
For example, our knowledge about the type, structure, and constituent mater-
ials of airplanes figures importantly in how we interpret the success of cubist
painting. In other words, to do so, we will employ a submetaphor of the
cubist painting is an airplane metaphor, which is the success of
cubist painting is the appropriate type, structure, and mater-
ials of the airplane.
This takes us to the level of mental spaces. What we find here is how the
potential frame-level knowledge is actually utilized in the painting. First, the
type of the airplane: what we can see is that the frame-level airplane is a very
light airplane. Second, the structure of the airplane: the airplane in the painting
looks like an early one from the experimental stages of airplane building, and it
looks like one which is quite “loosely” put together. Third, the materials: it is
made of cubist paintings (“The Violin” by Picasso).
These features are conceptually blended in the painting. Together they
portray an airplane that is hazardous for purposes of travel. The blend at the
mental spaces level seems to be coherent with Picasso’s and Braque’s initial
fears concerning their project. That is to say, at the level of mental spaces we
are dealing with a very specific conceptual metaphor that can be put as follows:
the fragility and uncertain fate of cubism at the time of its
emergence is the fragility and uncertainty that comes
from the lightness and constituent materials of the air-
plane that is about to take off.
The mental space level of the painting involves more. It also tries to answer the
question of who the leading figure of cubist painting is. It does so by making
use of the “operation” aspect of airplanes. We have the frame level knowledge
that the controller of the plane is the pilot. This maps onto the leading figure of
cubist painting, as represented in the Tansey painting: the pilot corresponds to
the leading figure. Thus we get a further mental space level metaphor: the
leading figure of cubism is the pilot of the plane (that is,
Picasso). Finally at the mental space level, the flying of an airplane frame is
also elaborated at this level: There is a second person (that is, Braque) on the
ground who is actively involved in the operation of the plane (running along
84 Domains, Schemas, Frames, or Spaces?
excitedly on the ground, thus possibly playing a less important role in the
operation of the plane). This reaffirms the person of the leading figure: Picasso
is the pilot and leading figure, while Braque is on the ground, figuring less
importantly in the movement.
Abstract Movement
But cubism is not only a complex abstract system, it is also an abstract
movement. Consider the following quote on the relevant sense of “movement”
from a dictionary: “b: a series of organized activities working toward an
objective; also: an organized effort to promote or attain an end - the civil rights
movement - a movement to increase the minimum wage” (from Merriam-
Webster Dictionary, www.merriam-webster.com/dictionary/movement).
At the image schema level the complex abstract system moves and moves
toward a goal. This gives us two image schemas in light of which we interpret
complex abstract systems, such as cubism, as moving: the movement
(activity) and source-path-goal (objective/purpose) schemas,
which can be represented as follows:
The two image schemas provide two image schematic metaphors: change
is motion and action is self-propelled motion.
More specifically, one level lower we find the following partial character-
ization of the domain of Cubism in Wikipedia: “Cubism is an early-20th-
century art movement which brought European painting and sculpture
historically forward toward 20th century Modern art. Cubism in its various
forms inspired related movements in literature and architecture. Cubism has
4.6 Visual Metaphors 85
been considered to be among the most influential art movements of the 20th
century.[1][2] The term is broadly used in association with a wide variety of
art produced in Paris (Montmartre, Montparnasse and Puteaux) during the
1910s and throughout the 1920s” (from Wikipedia, https://en.wikipedia.org/
wiki/Cubism).
Given the metaphor action is self-propelled motion from the Event
Structure metaphor (Lakoff 1993), we get an additional conceptual metaphor
that applies to cubism in particular:
purposeful artistic activities by a group of cubist artists is
organized travel by a group of travelers
At the frame level, the movement of cubist painting is an air journey,
resulting in the conceptual metaphor cubist painting is an air journey /
a flight. Given this frame level metaphor, the mappings can be laid out as
follows:
the air journey / flight ➔ the cubist painting movement
flying ➔ working / activities within the framework of cubism
the traveler in the air journey ➔ the painter in cubist painting
the destination of the air journey ➔ the goal of cubist painting
the beginning of the air journey ➔ the beginning of the cubist
painting
the end of the air journey ➔ the end of the cubist painting
These are fairly conventional mappings that are involved in any application of
the air journey source concept (more precisely, frame), not just cubism.
However, at this point we can ask why this mode of travel was (unconsciously)
chosen by the artist from among other modes of travel (such as car, train or sea
journey).
There are several answers to this question at the level of mental spaces. One
reason is that the success, viability – the functionality – of the movement was
not clear at the time of its inception. Success and functionality is metaphoric-
ally viewed as upward oriented; hence the metaphor functional is up, as
indicated by the idiomatic phrases up and running, get off the ground, take off.
Now, since action is motion, we can combine it with functional is up
to yield functional/successful action is upward motion, as in the
case of an air journey. Second, we can surmise that the choice is also facilitated
because the initial part of the journey is crucial in air travel (see, e.g., the many
metaphorical expressions concerned with the initial part of air travel: get off the
ground, take off, would not fly, launch). There is clear focus on the initial phase
of the movement in the painting; that is, the mapping that is emphasized is “the
beginning of the air journey ➔ the beginning of the cubist movement.” The
early history of the cubist movement was fraught with difficulties. The other
86 Domains, Schemas, Frames, or Spaces?
modes of travel would not have lent themselves to conceptually expressing this
uncertainty and excitement. Third, there might also be some cultural motiv-
ation for this choice, as we’ll see later.
The idea of cubism as an (abstract) movement manifests itself at the mental
space level in several further specific ways. A new, unconventional mapping is
introduced to the frame level metaphor. It is “the initiator / originator of the air
journey ➔ the initiator / originator of cubist painting.” In other words, there is
some focus on the initiators or originators of cubism: namely, Picasso and
Braque. This means that the originators are specified by name, whereas in the
frame level metaphor we only have the role of an unspecified traveler. Actu-
ally, in the painting one of the originators (Picasso) is also the traveler. Thus,
the roles of traveler and originator coincide in the painting (i.e., at the mental
space level). Moreover, in the painting one of the originators (Braque) is
depicted as remaining on the ground, thus receiving secondary status in the
movement. Finally, the painting also evokes the Wright brothers, who had
completed their first successful flight (in 1903) a few years before Cubism as a
movement actually started. The two artists, Picasso and Braque, thought of
themselves as the Wright brothers (Picasso often jokingly called Braque
Wilbur – the first name of one of the Wright brothers). The parallel brings
up for Picasso and Braque, and probably also for Tansey and the viewers of the
painting, the initial excitement associated with the two enterprises. In other
words, blending occurs at this level: the Wright brothers are conceptually fused
with Picasso and Braque.
This reading arises from historical knowledge – in the terminology of the
contextualist view of CMT, the conceptual-cognitive context. The recognition
of this parallel may have also functioned as a motivation for the choice of air
journey as a frame to conceptualize cubism by Mark Tansey.
Methodology
There are a number of different approaches to the study of metaphor within a
general CMT framework. These different methods sometimes compete with
one another and claim superior status in the field (on this, see Gibbs 2017b).
Lakoff and Johnson’s (1980, 1999) work was later seen as an “intuitive”
approach to the study of metaphor. The label was used because Lakoff and
Johnson and their followers were believed to rely on their own intuitions in
identifying metaphors and grouping them into sets of examples to demonstrate
the existence of conceptual metaphors. Corpus linguists argue that to find all or
most of the conceptual metaphors in a language, one needs to take advantage
of large corpora, such as the BNC (British National Corpus) or COCA (Corpus
of Contemporary American English) (see, e.g., Charteris-Black 2004; Deignan
2005; Stefanowitsch 2006). They emphatically point out that our linguistic
4.6 Visual Metaphors 87
intuitions are not adequate for studying metaphors. More recently, Kövecses
and his co-workers (Kövecses 2015b; Kövecses et al. 2019) proposed to revive
the older intuitive method. Following Kövecses’ (1986) initial suggestion,
I call the revised method the “lexical approach” and show that the lexical
method has advantages that the corpus-linguistic (more precisely, corpus-
based) method cannot ignore. Other metaphor scholars place emphasis on
the social-pragmatic uses of particular metaphorical expressions. They typic-
ally work with smaller databases than corpus linguists, and study the commu-
nicative functions of metaphors in authentic discourses (e.g., Cameron 2003;
Musolff 2006; Semino 2008). Some scholars pay attention to how frames and
metaphors are related and study various metaphorical constructions (see, e.g.,
Sullivan 2013). They often make use of the Framenet project that was
developed by Charles Fillmore.
Psycholinguists and cognitive psychologists study metaphors as a concep-
tual phenomenon and the embodied nature of metaphorical conceptualization
(see, e.g., Boroditsky 2001; Casasanto 2009; Gibbs 1994, 2006; Gibbs and
Colston 2012). This work is largely experimental – in the form of either “in
vitro” or “in vivo” experiments (see Kövecses 2005, on this distinction).
Neuroscientists take advantage of the most recent brain imaging techniques
and try to identify the neuronal activities that underlie the use of conceptual
metaphors in the brain (e.g., Bambini et al. 2011; Coulson 2008; Gallese and
Lakoff 2005 for an overview). Finally, groups of scientists are working to
build computational models of how humans use conceptual metaphors for a
variety of purposes (see, e.g., Feldman 2006; Narayanan 1999).
Thus, we have at least the following approaches dedicated to the study of
conceptual metaphors:
Intuitive approach
Corpus linguistic approach
Lexical approach
Discourse analytic approach
Framenet type approach
Psycholinguistic experimentation
Neuroscientific experimentation
Computational modeling
There are no doubt other approaches, but even this set shows very clearly the
variety of methods for the study of conceptual metaphor.
How can we pair off the different approaches with the different levels of
metaphor described above? The following pairing of approaches with the
levels is suggestive, rather than definitive. Most of the approaches can be used
to study several different levels. I suggest that the particular approaches apply
primarily to particular levels of metaphor, as shown in Figure 4.12.
88 Domains, Schemas, Frames, or Spaces?
--------------------------------------------------------------------------------------------------------
DOMAINS
FRAMES
– Psycholinguistics; Neuroscience
4.7 Conclusions
In the chapter I offered a four-level view of metaphor, which was demonstrated
with the building source concept. The suggestion was that the four levels
form a hierarchy in which conceptual metaphors occupy different levels of
schematicity, the highest level being that of image schemas and the lowest that
of mental spaces, with domains and frames in between. In the view proposed,
each conceptual metaphor is characterized by all four levels.
The mappings between the source and target concepts occur on the same
level: domains correspond to domains, frames to frames, and mental spaces to
mental spaces. In the case of the level of image schemas, the image schemas
structure highly schematic target concepts, such as complex abstract
systems, whose members are, for instance, the concepts of theory, mind,
nation, company, life, career, economy, and many others. The map-
pings at the various levels license metaphorically used words at different levels
(such as inside (image schema level), structure (domain level), and builder
(frame level), as shown above.
The four levels are connected vertically, in that the lower levels are more
specific versions of the higher, more schematic levels. Schematicity was
characterized, following Langacker (1987), as “relative precision of specifica-
tion” that applies to a concept.
Thus, conceptual metaphors occur at four levels of schematicity in an
interlocking vertical hierarchy of image schemas, domains, frames, and mental
spaces. In other words, conceptual metaphors cannot and should not be linked
to a single conceptual structure, such as frames or domains. Conceptual
metaphors are complexes of all four of these at the same time. I call this
structuring the “multilevel view” of metaphor.
The schematicity hierarchy is characterized by two major divisions. One is the
distinction between image schemas, on the one hand, and the other three levels,
on the other. While image schemas are analogue conceptual structures, the other
three are non-analogue ones (i.e., propositional) in nature. The other division
that was proposed is between levels that, on the one hand, belong to long-term
memory (image schemas, domains, frames), as opposed to the level of mental
spaces, that function online in working memory. Mental spaces activate higher-
level conceptual structures, while the latter structure mental spaces.
Insofar as the lexical approach to the study of metaphor can show, the
building source concept highlights two aspects in relation to the target
4.7 Conclusions 91
idioms but can also shed new light on how many idioms are organized around
a large “core” of high-level metaphors. Fourth, as regards how the multilevel
view of CMT applies to the analysis of visual metaphors, I examined a single
painting by American artist Mark Tansey from the perspective of extended
CMT. The multilevel view of metaphor proved especially helpful in account-
ing for the interpretation of the painting. Both the painter and the viewers
approach the work at a variety of levels of schematicity. Much of the concep-
tual materials on the image schema, domain, and frame levels are most
probably shared by the artist and the viewers. They are, to a large extent,
conventional and stored in long-term memory. It is on the mental space level
that the artist’s creative imagination can be captured. In a way, it can be
suggested that the painting is the expression of the conceptual product of the
artist’s relevant mental space in the course of producing the painting – without
the metaphorical processes that led to the final (static) product (i.e., the
painting). These are, however, recoverable from the product itself. Fifth and
finally, I discussed the issue of methodological “rivalry” in relation to the
question of what the “best” method is in the study of metaphors. Here, the
suggestion was that different methodologies and approaches apply to different
levels of metaphor and that they are all useful in doing their respective jobs.
All in all, it seems that by applying the multilevel view of metaphor we can
gain a great deal of coherence and uniformity in the description of seemingly
diverse phenomena in the study of metaphor.
5 Conceptual or Contextual?
metaphors. These experiences provide the common ground that allows con-
ceptualizers to produce and comprehend contextually induced, or generated,
metaphors in discourse.
The factors under cognitive styles, by contrast, reflect particular ways in
which experiences of the world need to be presented, given the prevalent
cognitive conventions and preferences of a language community. Such issues
as at which level of generality a metaphorical idea is presented (schematicity),
how it should be framed (framing), to what degree it should be conventional-
ized (conventionalization), which aspect of the body it should involve (experi-
ential focus), and others, are presentational in nature (for example, see
Kövecses 2015a). In general, the former set of factors (i.e., differential experi-
ence) respond to the question of “what” can prompt or prime the use of certain
metaphors, whereas the latter set to the question of “how” metaphorical
conceptualization needs to be presented in a language community. The latter
factors listed under differential cognitive style function as constraints on the
speaker-conceptualizer only, who is to follow the linguistic and cognitive
conventions of the language community. The hearer (conceptualizer 2/com-
prehender) is not constrained in this way.
Now let us turn to the various types of context and the more specific
contextual factors that constitute them. (For a detailed discussion and a number
of examples, see Kövecses 2015a).
Situational Context
The situational context comprises a variety of contextual factors. Most
commonly this type of context can be thought of as including the physical
environment, the social situation, and the cultural situation.
The physical environment can shape metaphorical meaning making. It
includes the flora, the fauna, the landscape, the temperature, the weather,
perceptual properties of the situation, and so on. For example, it is a common
observation that American English metaphors relating to the physical environ-
ment are characteristically different from those of other English-speaking
countries (see, e.g., Kövecses 2000c). The small-scale, local environment,
such as the visible events in or the perceptual properties of a situation, can
also make its influence felt in shaping metaphors.
The social situation consists of social aspects of life that typically center
around notions such as gender, class, politeness, work, education, social organ-
izations, social structure, and others. All of these can play a role in metaphorical
conceptualization. For example, Kolodny (1975, 1984) shows that American
men and women developed very different metaphorical images for what they
conceived of as the frontier in America in the period before the twentieth
century. While the women commonly thought of the American frontier as a
5.2 Types of Context and Kinds of Contextual Factors 97
Discourse Context
The discourse context involves the surrounding discourse, knowledge about
the main elements of discourse, the previous discourses on the same topic, and
the dominant forms of discourse related to a particular subject matter.
The surrounding discourse is simply the linguistic context – often referred
to as “cotext.” Viewed from the perspective of the discourse’s producer (the
speaker), elements of the preceding discourse (either by the speaker/concep-
tualizer 1 or the hearer/conceptualizer 2) can influence the (unconscious)
choice of metaphors, as was shown in an example taken from The Times by
Kövecses (2010b): “which helped to tilt the balance – and Mr Hain – over the
edge.” In this case, the contextually induced metaphor arises from the elliptical
use of the verb tilt in the phrase tilt Mr. Hain over the edge. It is the presence of
the word tilt in the immediate cotext that leads to the second use of the
metaphor.
Conceptualizers often rely on their knowledge concerning the main elements
of a discourse: the speaker, hearer, and the topic. For example, it occurs
commonly that in many newspaper articles knowledge about the topic as a
contextual factor leads to the creation of new metaphors (Kövecses 2010b,
2015a). An example of this involved David Beckham in an article in which the
journalist remarked: “Los Angeles Galaxy are sardines not sharks in the ocean
of footy” (Kövecses 2010b). Here the relevant knowledge includes that Beck-
ham played for the Los Angeles Galaxy soccer team and that Los Angeles is
located on the ocean with all kinds of fish in it.
The metaphors in one discourse can also derive from previous discourses on
the same topic. This can take a variety of forms ranging from elaborating,
98 Conceptual or Contextual?
Conceptual-Cognitive Context
This type of context includes the metaphorical conceptual system, ideology,
knowledge about past events, and interests and concerns.
Concepts can stand in a metaphorical relationship with one another (e.g.,
life is a journey, argument is war) in long-term memory. Given such
metaphorical relationships between concepts (such as between, say, argu-
ment and war), their presence or absence in the metaphorical conceptual
system may lead to the production and comprehension of particular metaphors.
(If the metaphorical connection does not exist in the user’s conceptual system,
he/she would not be able to come to use a metaphor.) A metaphorical concep-
tual system can function as context in this sense. Given an intended metaphor-
ical meaning (say, “supporting an argument”), we can search the conventional
metaphorical conceptual system for the best choice of metaphor. This happens
in cases where a conventionalized metaphorical meaning (“supporting an
argument”) is expressed via a conventional linguistic metaphor (such as
defend), with a matching target element activating the relevant mapping, or
conceptual correspondence, in an existing conceptual metaphor (e.g., the mean-
ing “supporting an argument” by means of the word defend in the argument
is war conceptual metaphor as based on the mapping “defending one’s physical
position in war corresponding to “supporting one’s position in an argument”).
The conventional metaphorical conceptual system is especially utilized this way
when there are no other overriding contextual factors in the discourse situation
that might influence the (unconscious) choice of a metaphor.
Ideology can also be a formative factor in how metaphors are used in
discourse. One’s ideological stance concerning major social and political
issues may govern the choice of metaphors (as work by, for instance, Goatly
5.2 Types of Context and Kinds of Contextual Factors 99
2007, shows). A good example of how ideology might influence the choice of
metaphors is George Lakoff’s (1996) study of American politics, where
conservatives tend to use the nation is a strict father family meta-
phor, while liberals prefer the nation is a nurturant parent family
version of the generic metaphor the nation is a family. For another
example, we can mention the Marxist version of the society is a building
metaphor with talk about “superstructure,” and so on, or the Marxist idea of
“class struggle.” Goatly (2007) is an important exploration into the metaphor-
based ideology of capitalism.
Being aware of past events and states (i.e., items in short-term and long-term
memory) shared by the conceptualizers may also lead to the emergence of
specific metaphors in discourse. A special case of this involves a situation in
which the speaker assumes that the hearer has a particular mental state. Such
memories of events can belong to the life of a community or an individual. It has
been often observed that the memory of historical events can lead to the produc-
tion (and comprehension) of some metaphors (see, e.g., Deignan 2003, Kövecses
2005). Different historical contexts can create differential preferences for par-
ticular life metaphors among Hungarians and Americans (see Kövecses
2005). The particular events in a specific communicative situation preceding
an act of metaphorical conceptualization may also produce similar effects.
People are commonly prompted to use particular metaphors (more pre-
cisely, metaphorical source domains) relative to their interests and concerns
about the world (see Kövecses 2005). Entire groups and individuals can be
said to have certain characteristic interests or concerns that may affect the way
they make meaning metaphorically. Since Americans are commonly regarded
as dynamically oriented, rather than passive, in their attitude to life, and,
relatedly, are sports-loving in general, it is not surprising that they use a large
number of sports metaphors. Similarly, if a person has some kind of profes-
sional interest, that person is likely to draw metaphors from his or her sphere
of interest (see Kövecses 2005). In other words, the kinds of activities we
engage in routinely affect the metaphors (the source domains) we use (see
Gelfand and McCusker 2001).
Bodily Context
A particular state of the body can produce particular metaphorical conceptual-
izations in specific cases, such as a poet’s or writer’s illness. Elsewhere,
I showed how Dickinson’s choice of metaphors may have been influenced
by her optical illness (see Kövecses 2010b). Such cases illustrate, more
broadly, that people’s bodily specificities influence which metaphors they tend
to use. There is experimental evidence (Casasanto 2009) that left-handers
prefer to use the moral is left, as opposed to the moral is right
100 Conceptual or Contextual?
conceptual metaphor. Such metaphors contrast with the metaphors that evolve
on the basis of the general properties of the human body (i.e., the correlation-
based primary metaphors – see Chapter 1).
Given this kind of evidence and the many metaphorical examples that are
based on local and temporary specificities of the human body (i.e., not on
universal embodiment), one might take the body as a further form of the
context. In this view, the body – especially those aspects of it that are activated
in the ongoing communicative situation – can influence the choice of meta-
phors in natural discourse (see Kövecses 2015a). The body is not only respon-
sible for the production of hundreds of conceptual metaphors through the
many correlations between subjective and sensory-motor experience (cf.
Grady, 1997a,b; Lakoff and Johnson 1999), but it can also prime the use of
particular metaphors in more immediate, local contexts (as further experimen-
tal evidence by, e.g, Boroditsky, 2001; Boroditsky and Ramscar, 2002; Gibbs
2006; Gibbs and Colston 2012 indicates). In other words, the body can lead to
the production of metaphors in discourse in the same way as the other
contextual factors previously mentioned can. Given this, we can think of it
as an additional context type.
L CO
OBA NT
EX
GL T
CON
CAL TE
situational LO XT discourse
context context
metaphor
used
conceptual-cognitive bodily
context context
Priming
It can be argued that (the knowledge and awareness of ) our experiences in the
local and global contexts can prompt the use of particular metaphors – either
conventional or novel ones. The various contextual factors can prime the use
of a metaphor in discourse. Priming is a well-studied cognitive process used
extensively in psychological and psycholinguistic experiments with a sizeable
literature (see, e.g., Boroditsky and Ramscar 2002; Casasanto 2009; Gibbs and
Colston 2012; and several other studies). Importantly, priming is based on the
simulation of some experience in the situational, discourse, bodily, and
conceptual-cognitive context. It is shared experience (the dynamically evolv-
ing common ground in a situation) that enables the production and comprehen-
sion of metaphors in discourse.
5.4 Operationalizing Contextual Influence on Metaphor Production 103
Let us take an example for this from work by Semino (2008), as described
by Kövecses (2015a). Semino studied the metaphors used by various partici-
pants at the 2005 G8 summit meeting in Scotland on the basis of an article
about the summit. The summit was accompanied by a major rock concert
called Live 8. Some participants assessed what the G8 summit had achieved
positively, while some had doubts concerning its results. Semino looked at one
such negative assessment she found in an article about the summit. She
comments in the following way:
In contrast, a representative of an anti-poverty group is quoted as negatively assessing
the G8 summit in comparison with the Live 8 concert via a metaphor to do with sound:
1.1 Dr Kumi Naidoo, from the anti-poverty lobby group G-Cap, said after “the roar”
produced by Live 8, the G8 had uttered “a whisper.”
The reference to “roar” could be a nonmetaphorical description of the sound made by
the crowd at the concert. However, the use of “whisper” in relation to the summit is
clearly a (negative) metaphorical description of the outcome of the discussions in terms
of a sound characterized by lack of loudness. Hence, the contrast in loudness between
the sounds indicated by “roar” and “whisper” is used metaphorically to establish a
contrast between the strength of feeling and commitment expressed by the concert
audiences and the lack of resolve and effectiveness shown by the G8 leaders. (Semino
2008: 3–4)
In my view, the metaphor whisper here emerges largely from the physical
(-social) context in which it is produced. Dr. Kumi Naidoo creates the meta-
phor whisper against a background in which there is a very loud concert and a
comparatively quiet summit meeting. We can think of the loudness and the
relative quiet of the occasion as perceptual features of the two events. Thus, the
original conceptualizer, Dr. Kumi Naidoo, chooses a perceptual property of
the physical context from all the experiential content that is available to him.
Conceptual Pathways
Out of the many potentially metaphor-producing experiences represented by
the various contextual factors above, a particular metaphorical meaning can be
conveyed if a specific target-domain meaning is expressible by a context-
induced source-domain meaning. This can be achieved if a particular concep-
tual pathway (made up of several conceptual metaphors and metonymies) can
be built between the two meanings (Kövecses 2015a). This way a particular
piece and kind of information (or experiential content) and, with it, a particular
context-induced metaphor is chosen out of the huge number of available
options in the situation.
A conceptual pathway is often a schematicity hierarchy, as discussed in
Chapter 4. The safety net, the in the mountain stage, and other examples show
104 Conceptual or Contextual?
and image schemas. With the creation of each new mental space, the context
changes dynamically, and it affects how a situation is conceptualized meta-
phorically in and through that mental space.
At the same time, of course, the repeated, long-term effect of a contextual
factor can lead to the (re)construction of a frame or domain that bears the
impact of that frame or domain (Figure 5.2). In other words, in a sense, frames
and domains are also context-dependent (not only mental spaces), but their
shaping by context occurs over an extended period of time (both in terms of
content and structure), unlike the online priming of mental spaces, which is
instantaneous.
All in all, then, we have the following picture:
analogue
structure
IMAGE SCHEMAS
…………………………………………………………………………
long-term memory
DOMAINS
non-analogue structure
FRAMES
------------------------------------------------------------------------------------
MENTAL SPACES
working memory
CONTEXT
The upward arrow indicates that it is at the level of mental spaces that
contextual influence on metaphorical conceptualization occurs.
The two highest levels, those of image schemas and domains, may be universal
(or at least widespread) in cultures when it comes to the metaphorical concep-
tualization of societies. But the frame level, and especially the lowest mental
spaces level are fairly culture- and even individual-specific. Sandburg captures
the idea of building a new American society through the idea of building a
skyscraper. At this level of conceptual metaphor, other poets may diverge
widely (though using the same conceptual metaphors at the higher levels),
depending on a large number of contextual factors (such as the physical
environment). In other words, we can conclude on the basis of this example
that elaborations of high-level conceptual metaphors may contribute to meta-
phor variation in poetry as well. Metaphors that may be universal or wide-
spread at higher levels of schematicity may turn into something that is specific
to particular poetic traditions or to individual poets.
The example demonstrates very clearly the usefulness of both the notion of
schematicity hierarchy and that of contextual influence in the production of
novel conceptual metaphors.
drawings, pictures, paintings, illustrations, statues, and many others. This variety
of visual metaphors is based on a genre-based classification of visual experience,
in that the forms of visual metaphors belong to different genres as convention-
ally distinguished. In them, we find conceptual metaphors that have been
discovered mostly on the basis of linguistic metaphors but also ones that are
specific to particular genres (see, e.g., El Refaie 2019). Fourth, the classification
of visual metaphors may also be based on their degree of conventionality. They
can be creative or structural (i.e., nonconventional or conventional). Fifth, visual
metaphors can be classified according to whether they employ a single modality
or several ones (monomodal vs. multimodal metaphors).
There are, however, other ways of classifying visual metaphors. Another
potential basis for the classification can be the degree of the sign-like nature of
the visual experience that constitutes the visual metaphors. My contention is
that visual metaphors can be composed not only of signs (i.e., sign-like visual
experiences), such as in advertisements, cartoons, and sculptures, but also of
visual experiences that we would not normally consider sign-like, such as
when we see a building or look at a scene in nature.
The suggestion builds on the idea that non-sign-like visual experiences
constituting visual metaphors utilize the same processing mechanisms as the
sign-like ones. The interpretation and creation of both kinds of metaphorical
visual experience relies on framing, mapping, recognizing correlations between
experiences, finding resemblance between experiences, creating new experiences
represented by new frames, getting primed by experience, resorting to already
existing metaphors in long-term memory, and several others. In other words,
conceptual metaphors realized visually are not to be found outside the human
organism; they are found internally as human beings process experiences, either
sign-like or non-sign-like, by means of these conceptual mechanisms.
Here, the winner of a competition occupies the highest position at the award
ceremony because he or she has social control over the other competitors. That
is, the person with the most power or control is placed highest. Social power or
control is reflected in higher physical position.
There are many other social situations where we can see the conceptual
metaphor at work. For example, the king sits or stands higher than his subor-
dinates, the speaker of the house sits higher than the representatives, teachers
often stand on a podium in classrooms, and so on. In general, elevated
platforms of various sorts, besides their practical functions, serve the purpose
of indicating how some people have more power than others.
In many other cases it may not be immediately obvious that the metaphor
control is up is present and it metaphorically constitutes certain aspects of
reality (Figure 5.4). By way of an example, imagine that you are a foreign
tourist in a city and you encounter the following visual image:
110 Conceptual or Contextual?
The image is about Budapest and it shows the Buda Castle, the former Royal
Palace, on top of the castle hill on the Buda side of the river Danube, which
separates the older part of the city, Buda, from the newer part, Pest.
The image we can see above is the photograph of a “real” visual image that
a foreign tourist (or, as a matter of fact, any other person) can see in Budapest.
Is the real image (perceived by a tourist) and/or the photograph of the image
(shown above) in any sense metaphorical? And, importantly, how can we
find out?
It should be noticed that the image does not portray people who could be
said to be in some obvious power-based relationship with one another. We can
only see buildings, a hill, the river, boats, etc. And yet, interestingly, foreign
visitors to Budapest can interpret what they see via the control is up
metaphor.
In a small and informal research project, I investigated the metaphors that
foreign tourists who visit Budapest use to interpret their experiences (see
Kövecses in press). I did a simple Google search to find descriptions of
people’s experiences and impressions of Budapest. I only gathered descriptive
statements that contained metaphorical expressions. This way, a database of
about 130 pages long was compiled. After eliminating duplicates and the
repeated uses of the same metaphor, I ended up with an 83-page-long list of
5.8 Two Kinds of Visual Metaphor 111
power and control. But in addition, the sight of the river by itself can activate
in its viewers such metaphors as powerful action is forceful motion
(coursing with momentum, boats roared through the waters, with grace and
power). This metaphor derives from the primary metaphor action is motion.
Moreover, the entire scene also involves the good is light or bright
metaphor, as shown by the sun’s rays reflecting from the surface of the river.
This is a more or less literal description of what he can see around him from the
window of a seaside house (Figure 5.6). But when the tide goes out he (they)
may see something like this:
114 Conceptual or Contextual?
The visual scene consists of low tide at Dover Beach with the sea withdraw-
ing, which gives him the following idea:
The sea of Faith
Was once, too, at the full, and round earth’s shore
Lay like the folds of a bright girdle furled.
But now I only hear
Its melancholy, long, withdrawing roar,
Retreating, to the breath
Of the night-wind, down the vast edges drear
And naked shingles of the world.
(www.artofeurope.com/arnold/arn1.htm)
The stanza may be interpreted in the following way: The sea that covers the shore
gives protection to the earth, but when it withdraws at low tide, it leaves the earth
exposed. The same can happen to faith. Its importance for people may diminish,
but when it does, it leaves people unprotected. In other words, the sea corresponds
to faith, the earth corresponds to the people, and the covering of the earth by the
sea corresponds to the protection of the people by faith. Thus, the visual image of
the low tide can trigger thoughts in the poet, which is based on some skeletal
structure shared by the withdrawing of the tide and the diminishing of faith.
However, this recognition of shared schematic structure can also occur in
reverse order. It may well be that the poet had these ideas in mind about
5.9 Conclusions 115
Christian faith, and then he found the appropriate visual image to express it. As
a matter of fact, in all probability the two can also co-occur simultaneously.
But this is not the crucial point. What is important is that creative, unconven-
tional resemblance metaphors can emerge as a result of visual images of the
real world and can combine with abstract ideas.
In this section, I noted that, in addition to visual aspects of many metaphor-
ical source domains (as in, e.g., building that is found in the metaphor
theories are buildings), there are visual images (e.g., the sight of Buda
Castle or Dover Beach at low tide) in the online physical context of metaphor-
ical conceptualization that can serve as source domains of novel metaphors.
I suggested that we can think of such cases as the visual images priming the
selection of metaphorical source domains.
In Chapter 4 and the present one, I examined a variety of visual experiences
that appear to evoke visual metaphors. This is a range of experience types that
extends from sign-like visual experiences (like a painting) to non-sign-like
visual experiences (like a visual image of an artifact or nature or a city scene).
Visual metaphors are evoked by paintings through winner’s podiums all the
way to cityscapes and scenes in nature. The last two (non-sign-like) cases,
cityscapes and natural scenes, are not commonly subjected to serious examin-
ation from a CMT perspective. However, they provide us with a new challenge
in the study of visual metaphors, in that they greatly extend the range of visual
experiences that might give rise to visual metaphors.
At the same time, remarkably, the comprehension or interpretation of all of
these visual experiences, including sign-like and non-sign-like ones alike,
makes use of the same metaphorical processing mechanisms. The visual
metaphors that are evoked by visual experiences can be based either on
correlations or resemblance. This latter bifurcation applies equally to non-
sign-like visual experiences, such as cityscapes and scenes of nature.
5.9 Conclusions
In the chapter, I argued for a broad conception of context in metaphorical
conceptualization – one that covers our cognitive interaction with various
elements and properties of the situation of discourse, the discourse itself, the
conceptual-cognitive background, and the body of the speaker and hearer.
These dimensions of our interaction with the world were labeled the situational
context, discourse context, conceptual-cognitive context, and the bodily con-
text. All of these context types and the specific contextual factors that make
them up can influence the creation of metaphors in discourse.
This view of metaphor is, I feel, preferable to other views within cognitive
linguistics because it extends the study of metaphor beyond those that are
body-based in the usual sense in conceptual metaphor theory (i.e., universal
116 Conceptual or Contextual?
One of the major criticisms of conceptual metaphor theory has been that it is
not capable of accounting for meaning in actual occurrences of metaphorical
language in real discourse (see, e.g., Cameron 2003; Deignan 2005; Goatly
1997; Semino 2008; Steen 2007, and many others). This kind of criticism
makes perfect sense if we consider the multilevel view I am outlining in this
book. In the “standard” view of CMT, researchers work on the levels of image
schema, domain, and frame. These are conceptual structures that are decon-
textualized patterns in long-term memory that can account for metaphorical
meaning in the most general ways. On the one hand, image schemas can
explain why certain metaphors can be meaningful through embodiment and,
on the other, the conceptual metaphors on these levels (image schema, domain,
and frame levels) can account for metaphorical meaning through the mappings
that constitute the conceptual metaphors and the inferences that can be made
on the basis of such mappings. However, both of these conceptual mechanisms
(providing meaningfulness and accounting for decontextualized meanings) are
fairly general ones that cannot capture the contextually variable richness of
metaphorical meaning in naturally occurring discourse.
What makes it possible to get at this richer and more specific aspect of
metaphorical meaning, but at the same time preserve these general aspects of
meaning, is to think of metaphor as a multilevel phenomenon that occurs also
at the level of mental spaces. As I have noted in previous chapters, online
metaphorical activity necessarily makes use of the conceptual structure of
mental spaces. This is the level where we produce and comprehend metaphors
in natural discourse. Specifically, it is at this level that speakers
use metaphors online (i.e., produce and comprehend metaphors)
use them in a fully contextualized way
use them with specific socio-pragmatic functions
add emotional value to them
create novel metaphors
use metaphors deliberately
create metaphors as a result of contextual priming
117
118 Offline or Online?
mention of the same person’s freezing has a role to play in this. Freezing is the
opposite of body heat that is associated with anger and that gives rise to heat-
related anger metaphors. The image of a hot stove is evoked as a result of the
mental-space level interaction of the being cold frame and the body heat
frame, in that a hot stove can be used to prevent someone from being cold.
A hot stove is a device that can be a part of the being cold frame but also of
the body heat frame that serves as the basis for the anger is heat
conceptual metaphor. In addition, a stove is a container just like the angry
person’s body is a container. In sum, at the level of mental spaces an aspect of
the discourse (literally freezing) can prime the speaker to introduce a non-
conventional metaphorical expression that also matches an established con-
ceptual metaphor (anger is heat).
Edwards makes the point that this particular use of “boiling” establishes
accountability and provides justification for one’s (the speaker’s) actions.
The passage is part of an interview in a counseling session, in which Jimmy
talks about how his wife’s “provocations” during an evening in a pub led him
to leave her in the pub (with the man she was “messing” with) and to go home.
The use of “boiling” shows that the intensity of his anger matches the intensity
of the “provocations,” thus making his wife accountable for his anger. At the
same time, it also shows that the intensity of his anger provides justification for
leaving the pub, go home, and, later, even assault his wife. “Boiling anger”
presents him as being passive in relation to events in the sense that he appears
to believe that such a high intensity of anger justifiably leads to those actions
on his part. Clearly, the use of “boiling” here goes far beyond the expression of
some metaphorical content. It fulfills extremely important narrative-discursive
functions that constitute another and indispensable part of the expression’s
meaning in a particular context.
relationship between (A) and (BCDE) and (B) and (CDE) there is no pressure
to resolve the incompatibility in imagery because of the grammatical structure
of the discourse, in the case of (CDE), which is a relationship between a single
clause, there is a pressure to resolve imagistic incompatibility between “tide,”
“red tape,” and “choking Britain.” Given such examples, Kimmel draws the
conclusion: “Only the close syntactic integration of two metaphors within a
clause can enforce or foster a close integration of their semantic content qua
imagery. Where mixed metaphors occur across clauses no ontological clashes
will be felt to begin with and secondary mechanisms to keep the clash at bay
are dispensable. If this is correct, making sense of mixed metaphors is a natural
byproduct of default clause processing” (Kimmel 2010: 110).
This provides a potential explanation of why most mixed metaphors are not
taken to be cases of incompatible images and are, thus, not problematic for
metaphor interpretation.
However, convincing as far as it goes, Kimmel’s account does not explain a
number of issues relating to mixed metaphors. These include:
(1) Why are imagistically incompatible metaphors selected at a particular
point in discourse?
(2) Why are mixed metaphors so common?
(3) Why do we have cases of metaphorically entirely homogeneous
discourse?
(4) Why do we often have widely divergent source domains inserted into the
discourse?
(5) How do we comprehend mixed metaphors?
Let us briefly discuss each of the questions above.
many different aspects to them, and these aspects normally require different
source domains, or frames, for their conceptualization, as was noted above.
Assume, for example, that the topic of discussion is the family. The notion
(domain or frame) of family comes with a large number of different aspects
(constituted by conceptual elements), including parents, children, the raising of
children, the cost of raising children, the family as a cohesive group of people,
and a host of others. Below is a sentence I found about discussing some of
these aspects, as cited by David Brooks in the International Herald Tribune
(November 17–18, 2012): “The surest way people bind themselves is through
the family.” (my bolding, ZK) What does way have to do with bind? The word
way belongs to the motion frame, while bind to the physical joining
frame (inside the frame a group of physically connected entities).
These are incongruent images (i.e., metaphorical source domains) that create
mixed metaphors.
To account for why these and many other incongruent images are so
common, I propose, in line with the argument in the previous subsection, that
all target domain concepts that constitute the topic (i.e., subject matter) of
discourse consist of a number of different aspects, and that each of these
aspects can be conceptualized metaphorically by means of different source
domains – domains that may yield incongruent images when juxtaposed. To
take the example above, as the discussion of the family domain is unfolding,
various aspects of it are mentioned; for example, the issue of what kind of
social unity is provided by the family and the issue of how a certain state can
be achieved (“The surest way people bind themselves is through the family.”).
As regards the second metaphor of the sentence (bind), it is based on the
conceptual metaphor cohesive social groups are groups of physic-
ally connected entities. Given this metaphor, one of its mappings “the
physical joining of entities ➔ the formation of social groups” accounts for the
use of bind. Social groups constitute more or less stable states of intercon-
nected entities. The issue of how more stable states are achieved is another
aspect of the family/social group domain. And since the issue of how certain
states are achieved in general is metaphorically conceptualized as manner
(of action) is path (of motion), which is a submetaphor of (deliber-
ate) change of state is (forced) motion, we can account for the use
of the word way in the sentence above. The resulting images are fairly
incongruent when juxtaposed (way – path / motion vs. bind – the phys-
ical joining of entities / a group of physically connected
entities), but their incongruence is natural since we use different source
domains to conceptualize different aspects of the target domain, as predicted
by the notion of meaning focus. As pointed out by Kimmel (2010), this is very
common in natural discourse; it seems to be the dominant pattern – the rule,
rather than the exception.
6.2 Mixing Metaphors 127
Performance targets are identical to the puissance at the Horse of the Year Show. You
know the one – the high-jump competition, where the poor, dumb horse is brought into
the ring, asked to clear a massive red wall, and as a reward for its heroic effort is
promptly brought back and asked to do it all over again, only higher.
I’ve never felt anything but admiration for those puissance horses which, not so dumb at
all, swiftly realize that the game is a bogey. Why on earth should they bother straining
heart, sinew and bone to leap higher than their own heads, only to be required to jump
even higher? And then possibly higher still.
Hard work and willingness, ponders the clever horse as he chomps in the stable that
night, clearly bring only punishment. And so next time he’s asked to canter up to the big
red wall, he plants his front feet in the ground and shakes his head. And says, what do
you take me for – an idiot? (Melanie Reid, The Times, Monday, February 4, 2008)
In this case, the same conceptual metaphor, life is a horse show, lends
coherence to a single discourse. The novel conceptual metaphor that structures
the text is characterized by a number of mappings: puissance horses are compared
to people, riders to managers, the red walls as obstacles to the targets people have
to achieve, having to jump over the obstacles to being subject to assessment,
clearing the obstacles to achieving the targets, raising the obstacles to giving
more difficult targets, the horse show to life, and so on and so forth. This
elaborate metaphorical analogy provides a great deal of structure for the text.
However, it may be even more important than the coherent organization of
the discourse that the author wishes to make a point emphatically and vividly
by the use of the homogeneous metaphor. This expressive function of the
discourse can often provide an explanation for the use of a single coherent
metaphorical image in discourse.
In addition to its organizational and expressive functions, metaphorically
homogeneous discourses can also be used for didactic purposes. Parables are
commonly used this way. Consider the parable of the sower in Matthew 13
(Biblegateway, www.biblegateway.com/passage/?search=Matthew+13&ver
sion=NIV):
128 Offline or Online?
1
That same day Jesus went out of the house and sat by the lake. 2 Such large crowds
gathered around him that he got into a boat and sat in it, while all the people stood on
the shore. 3 Then he told them many things in parables, saying: “A farmer went out to
sow his seed. 4 As he was scattering the seed, some fell along the path, and the birds
came and ate it up. 5 Some fell on rocky places, where it did not have much soil. It
sprang up quickly, because the soil was shallow. 6 But when the sun came up, the plants
were scorched, and they withered because they had no root. 7 Other seed fell among
thorns, which grew up and choked the plants. 8 Still other seed fell on good soil, where
it produced a crop – a hundred, sixty or thirty times what was sown. 9 Whoever has ears,
let them hear.”
Here Jesus teaches people about the kingdom of God. In doing so, he uses the
conceptual metaphor god’s words are seeds. By using a consistent image,
he feels he can convey his message more clearly to people.
Characteristic of the discourses structured by the life is a horse show
and god’s words are seeds metaphors is their conceptual homogeneity.
This is provided for by consistently applying images (elements: ideas, mean-
ings) from the same source to structure the various aspects of the target. In
such cases, as a result of the consistent structuring (based on a single source
image), no or very few mixed metaphors are found in discourse. What
discourses such as the above show is that certain discourse functions (vividly
making a point, teaching people) may provide fairly strong motivation for the
use of a single consistent metaphorical image (unlike the typical case of
metaphorically varied imagery, where the discourse has less marked rhetorical
or pragmatic functions).
metaphors often provide images that may not fit the conventionally available
source domains of a given target; nevertheless, they are easily understood
through the supportive role of context. Through the notion of contextual
priming that was discussed in Chapter 5 we can account for the use of widely
divergent metaphors used in some instances of natural discourse.
Kövecses 2002/2010; Lakoff 1993; Lakoff and Johnson 1980, 1999) should
give way to the “network model” that comprises several input spaces, a generic
space, and the blended space.
More recently, Fauconnier and Turner (2008: 53) took this argument one
step further as we can see in the passage below: “What we have come to call
“conceptual metaphors,” like time is money or time is space, turn out to
be mental constructions involving many spaces and many mappings in elabor-
ate integration networks constructed by means of overarching general prin-
ciples. These integration networks are far richer than the bundles of pairwise
bindings considered in recent theories of metaphor.”
In other words, the suggestion is that what are known as the typical
examples of conceptual metaphors are not simply “pairwise bindings” but
are the products of “elaborate integration networks.” All conceptual metaphors
are in fact blends.
Most generally and significantly, CIT researchers claim that metaphors,
together with analogy, framing, metonymy, etc., are surface cognitive phe-
nomena that “are consequences of the same basic human ability for double-
scope blending” (Fauconnier and Turner 2008: 54). In other words, metaphors
are relatively surface products of the deep, underlying process of conceptual
integration.
I agree with many of these suggestions mentioned above. For example,
I agree that in many cases CMT should move beyond structure-mapping and
inference transfer. I also accept the criticism that in many cases mappings
occur not only in a source-to-target direction. And most obviously, I also think
that blended spaces are clearly needed to account for a variety of examples that
CMT alone could not account for.
Nonetheless, in this section, I offer a somewhat different view on the
relationship between CMT and CIT than what we saw above. I group what
I have to say about the relationship around three topics. In particular, the first
of these concerns the role and importance of structure mapping both in CMT
and CIT. The second has to do with the claim that blending is somehow a more
basic, deeper cognitive operation than metaphor. The third is the issue whether
blending is indeed a more complex cognitive phenomenon than metaphor. My
answer to each of these questions will be a “partly yes, partly no” response.
GENERIC SPACE
A person
Employing a sharp tool to a body for a
purpose
BLEND
some underlying similarities given in the generic space: In both inputs, there is
a person who employs a sharp tool to a body for some purpose. This is
represented in Figure 6.1 above, where the entire CIT network is given, but
for now we only focus on the notion of the schematic knowledge shared by the
input spaces.
In other words, CIT spells out the similarities between source and target at a
schematic level. The similarities form the basis for bringing together the two
domains in the network. According to CIT scholars, this allows conceptuali-
zers/speakers to set up correspondences, or mappings, between elements of the
domains. That is, given the generic space above, we can map the butcher onto
surgeon, the cleaver onto the scalpel, etc.
136 Offline or Online?
1980). Strictly speaking, resemblance is not a truly viable option to explain the
problem, since the two concepts must already exist before any similarity is
perceived between them (see above).
Correlation in experience, however, is more promising. In primary metaphors
such as purposes are destinations, the destination can be generalized
into a purpose concept: the desire to get to a physical location can lead to the
emergence of the notion of purpose that is not limited to getting to a location
but applies generally (on this process, see Kövecses 2013 and Chapter 3.)
Another, and less recognized, way for abstract concepts to emerge is
through schematization. In several publications (see Kövecses 2011c,
2015c), I argue that metaphors such as heaven is an ideal physical
place, causation is progeneration, and providence is nurtur-
ance are based on the source domain schematized into the target. This is
happening via a metonymic process. The specific rich imagery of the source is
bleached out to yield such abstract (target) domains as heaven, causation,
and providence.
opponent 1 / self
opponent 2 / emotion
struggle / attempt at control
etc.
However, I do not think that this would yield a substantially different result
from the suggestion made by CMT. These are cases where two frames have the
same schematic structural relations between the elements, and so the elements
in the frames are compatible with each other in a straightforward manner. In
the CIT account, the mappings from source-to-target are simply renamed as
blended counterparts in the blend. It is not clear whether the idea of “maps
onto” and “are blended/integrated/fused” mean something very different in
such cases. It can thus be suggested that the two accounts are simply termino-
logical variants.
True cases of blending arise when, over and above the compatibilities, there
is some incompatibility between the two frames. Consider the well-known
example offered by Turner and Fauconnier (2000) and briefly commented on
above in the chapter, God, he was so mad smoke was coming out of his ears.
The sentence is based on the anger is heat conceptual metaphor (in its
two versions: hot fluid in a container or fire). There are several
mappings between the angry person and heat (either boiling pot or fire) that
obtain between the compatible elements in the two domains:
the container ➔ the angry person’s body
the heat ➔ the anger
the degree of the heat (of fluid) ➔ the intensity (of anger)
the cause of the heat ➔ the cause of anger
etc.
These elements in both frames constitute a structured whole. The two struc-
tures are compatible with each other, just like in the emotion is an oppon-
ent metaphor we saw above. However, in the “smoke was coming out of his
ears” example, there is a major incompatibility. Here, incompatible elements in
the source and target are brought together to form a conceptual unit that is not
part of either the source or the target: The element that indicates a high degree
of heat (the smoke or fume or steam) is conceptually integrated with an
element of the target – the angry person’s body (more specifically, the head).
The result is a smoking, fuming, or steaming head (of the angry person). The
sentence creates a new frame (a blended space) in which smoke/steam is
coming out of the person’s ears. The new frame results from merging two
otherwise incompatible elements from the source and target frames.
But there are even more complicated cases that we can account for by
making use of the idea of blending as arising from some incompatibility
6.3 Conceptual Metaphor and Conceptual Integration 139
between the elements of the source and target. Take the surgery is butch-
ery metaphorical blend discussed previously. We noted above that much of
the conceptualization involved in this example is very much like a straightfor-
ward set of metaphorical mappings: certain elements of the source correspond
to certain elements of the target. Here, we have a systematically organized
frame with elements, butchery, mapping onto the elements of another
systematically organized frame, surgery.
As Figure 6.1 above shows, on the blending view, in addition to the two
input spaces of butchery and surgery that are connected by a set of
mappings, or correspondences, as above (except for the last correspondence),
we have a generic space in which there is a person who employs a sharp tool to
a body for a purpose. There is also a blended space. This space inherits from
the source input the butcher and the means of butchery and from the target
input the surgeon, the patient, some tool, the operating room, and the goal of
healing. Thus, in the blend there is a surgeon in the role of a butcher who uses a
tool and the means of butchery for the purpose of healing a patient. But, of
course, the surgeon who uses the means of butchery cannot do a good job in
trying to heal a human patient. The blend set up this way leads to the
interpretation of the surgeon as being ineffective, nonprofessional, sloppy,
careless, etc., and, ultimately, incompetent.
Kövecses (2011b) uses the idea of “meaning focus” to analyze the example
“This surgeon is a butcher.” By employing the notion of meaning focus (see
Kövecses 2000a, 2002/2010), we could eliminate the problem associated with
“standard” conceptual metaphor theory: the problem that, on that analysis,
there is no account of why the feature of sloppiness or carelessness (or
incompetence) is mapped onto the surgeon. The view based on the meaning
focus of the source domain would maintain that the feature is mapped onto the
target because it is one of the meaning foci associated with butchers, as
indicated by the various senses of the word as defined by Merriam-Webster:
1 a: a person who slaughters animals or dresses their flesh b: a dealer in meat
2: one that kills ruthlessly or brutally
3: one that bungles or botches
4: a vendor especially on trains or in theaters
Sense 3 clearly indicates that butchers are regarded as inherently incompetent.
Given this conventionalized sense of the word and given that source domains
map their main meaning focus (whose selection from several potential foci
may depend on the context) on the target, we can understand why the meta-
phorical statement means what it does.
In addition, to account for the meaning of the sentence, we also need some
metonymies. There is a metonymic relationship between the category as a
whole and the property as a part, yielding the metonymy category for
140 Offline or Online?
its property (see Kövecses and Radden 1998). That is, the word butcher is
used in the sentence to metonymically indicate incompetence.
Let us return to the issue of the butcher being “careless, sloppy, imprecise,”
that is, incompetent. Butchers are not inherently incompetent, but sometimes
we take them to be such. Why? The reason may be that the actions performed
by the butcher appear to be sloppy in contrast to the surgeon. We compare the
butcher’s actions with the “precise” and “refined” actions of the surgeon (cf.
the phrase “with surgical precision”). In other words, we interpret how the
butcher works in reference to the surgeon’s work, that is, with the surgery
frame in the background. Therefore, we extend the primary meaning of the
word butcher (“who slaughters animals and dresses their flesh”) to “incompe-
tent.” This newly derived meaning will then be projected to and will charac-
terize the particular surgeon as well. We can think of the projection as a case of
conceptual integration. The projection goes to a new space, the blended space,
where the “careless, sloppy work” of the butcher replaces the “precise and
refined work” of the surgeon. In this way, the blend contains mostly infor-
mation that the surgery frame contains, with the major difference that the
particular surgeon will here be regarded as doing “careless and sloppy work”
and, hence, “being incompetent.” The surgeon in the blend assumes the
contextually defined new meaning focus of the butcher. We can diagram this
as in Figure 6.2 below on the next page.
The above solution to the question of how the meaning of the sentence
arises is different from the CIT solution proposed by Grady et al. (1999). In
CIT, the essential elements of the blend were the means of butchery and the
goal of surgery, as well as the conflict between the two, leading to the property
of “incompetence.” In the present suggestion, the property of “incompetence”
is projected into the blend from the input space of butchery. The property of
“incompetence” emerges in the concept of butchery in light of and against
the background of the concept of surgery, and then it is projected into the
blend, in which the property will now characterize the surgeon. That is, critical
conceptual material in the blend will derive from the source. Thus, if the
contextually defined meaning focus of the source domain of butcher is that
he does careless, sloppy work (relative to the surgeon), it will carry over this
property to the target (and to the blend, as above).
Let us now turn to a case where the incompatibility is between two stories,
as in parables. In some parables, when a new didactic point is made by the
speaker, the two stories involved in the parable are incompatible, and despite
the several systematic correspondences between the source and the target, the
source story does not match the target story in certain respects. This may cause
surprise or difficulty in interpretation (see the section on “mixing metaphors”
above). The Parable of the Workers in the Vineyard in the New Testament of
the Bible is a good example of such a situation:
6.3 Conceptual Metaphor and Conceptual Integration 141
GENERIC SPACE
A person
20 ”For the kingdom of heaven is like a landowner who went out early in the morning
to hire workers for his vineyard. 2 He agreed to pay them a denarius[a] for the day and
sent them into his vineyard.
3
”About nine in the morning he went out and saw others standing in the marketplace
doing nothing. 4 He told them, ‘You also go and work in my vineyard, and I will pay
you whatever is right.’ 5 So they went.
“He went out again about noon and about three in the afternoon and did the same thing.
6
About five in the afternoon he went out and found still others standing around. He
asked them, ‘Why have you been standing here all day long doing nothing?’
7
”‘Because no one has hired us,’ they answered.
“He said to them, ‘You also go and work in my vineyard.’
8
”When evening came, the owner of the vineyard said to his foreman, ‘Call the workers
and pay them their wages, beginning with the last ones hired and going on to the first.’
142 Offline or Online?
9
”The workers who were hired about five in the afternoon came and each received a
denarius. 10 So when those came who were hired first, they expected to receive more.
But each one of them also received a denarius. 11 When they received it, they began to
grumble against the landowner. 12 ’These who were hired last worked only one hour,’
they said, ‘and you have made them equal to us who have borne the burden of the work
and the heat of the day.’
13
”But he answered one of them, ‘I am not being unfair to you, friend. Didn’t you agree
to work for a denarius? 14 Take your pay and go. I want to give the one who was hired
last the same as I gave you. 15 Don’t I have the right to do what I want with my own
money? Or are you envious because I am generous?’
16
”So the last will be first, and the first will be last.” (Matthew 20:1–16, New
International Version)
Most people, both in Jesus’ time and today, are surprised, and even find it
unfair, that Jesus gives the same pay to those who worked the whole day and
those who worked only one hour. This can happen because Jesus changes the
source domain story according to the target story. The target story is about the
kingdom of heaven, where all people are invited (just like the workers by the
landowner) and receive the same “pay.” However, the pay is not financial
reward but forgiveness and mercy that all people who ask for it will get in
equal amount. In other words, the kingdom of heaven is only partially similar
to our world, and Jesus adjusts the source story according to how things work
in the kingdom of heaven, the target story.
I suggest that the cognitive process that produces such an interpretation is
conceptual integration (see, e.g., Turner 1996). In blending, two partially
similar, thus essentially compatible, domains are merged, rather than the
source simply structuring the target. In my view, this merging is possible
and arises when there is some incompatibility between the (otherwise similar)
source and target. The incompatible element in the target (i.e., the element that
does not match an element in the corresponding source) will figure importantly
in the intended interpretation of the parable. In more specific terms this means
that the speaker will use elements of the source that are not coherent with the
rest of the source but that fit the target. Technically, this is achieved by the
speaker projecting the target-compatible (but source-incompatible) elements
into a blended conceptual space, where the target-compatible elements (every-
one gets the same “amount” of forgiveness and mercy) are fused with the
source-incompatible ones (everyone gets the same wage, regardless of how
many hours they work). Thus, in the blend all workers get the same amount of
forgiveness and mercy, no matter how long they have worked for it (i.e., when
they requested it).
In sum, blending is needed when incompatible elements between two
otherwise partially compatible domains are used in meaning construction.
6.3 Conceptual Metaphor and Conceptual Integration 143
a fourth it is drown, and so forth. The answer again has to do with context. On
the CIT analysis, for example in the case of “Cowboys Corral Buffalos,” in the
blend we would have a university team blended with an opponent and
cowboys, another university team, blended with the other opponent and
buffalos, and, finally, the action of defeating with corralling. But, again, why
is corral used out of many other options? The reason seems to be that it is the
semantic content, or meaning, of the immediate linguistic context (and not its
phonological shape, unlike the previous example above) that facilitates the
choice of a metaphorical expression: in the Wild West frame, cowboys are in
the business of corralling animals. The same kind of explanation applies to the
other cases.
Third, Coulson and Oakley (2003) examine still another headline: “Coke
Flows Past Forecasts: Soft Drink Company Posts Gains” from USA Today.
They suggest that the literal predication of the drink Coca Cola (namely, that it
flows) and the metaphoric predication of flow for the Coca Cola Corporation’s
profit are blended in the blended space. In this case, we seem to have double
motivation for the use of flow in the network. One is of course the immediate
linguistic context with the word coke, which has the property that it flows. The
other is the general topic of the discourse, Coca Cola, which is a liquid. Thus,
here, there is both the immediate linguistic context and the general topic, or
subject matter, of the discourse (a liquid) that is at work in motivating the
particular blend that is created.
Fourth, let us take one of the best known examples of blending: “If Clinton
Were the Titanic, the Iceberg Would Sink” (Turner and Fauconnier 2000).
This is a complex example whose analysis by Turner and Fauconnier shows
the full power of CIT. What their analysis does not show in an explicit and
systematic way, however, is why we have the particular input spaces that we
do in the network of spaces. To explain this, we need a model of meaning
making that takes seriously and includes a theory of contextual influence on
the construction of meaning – both for metaphor and blending (see Kövecses
2015a). Actually, Turner and Fauconnier do mention that this particular
blend could not have arisen in 1992 without the Titanic catastrophe to begin
with and without the wide popularity of the movie inside the Washington
Beltway during the Clinton presidency. However, this somewhat anecdotal
comment does not substitute for a full theory of contextual influence on
meaning construction. Therefore, I would add to the comment made by
Turner and Fauconnier that people often produce metaphorical blends by
relying on “context models” (on this, see Van Dijk 2009) that help them
activate aspects of context (i.e., contextual factors), which results in introdu-
cing an input space (typically a source concept) into the network (see
Kövecses 2015a).
6.3 Conceptual Metaphor and Conceptual Integration 145
Some Issues
There appears to be some rivalry between CMT and CIT. Many scholars
believe that they have to choose one or the other as their research tool because
one is somehow better than the other. I believe that the conclusions that we can
draw from the preceding analyses defuse this kind of tension: there is abso-
lutely no reason to think that CMT is better or that CIT is better. They very
much depend on and complement each other. One without the other would be
much weaker. This conclusion is shared by Grady et al. (1999: 120), who state:
“. . . we propose that because they tackle different aspects of metaphorical
conceptualization, the two frameworks are largely complementary. The con-
ventional conceptual pairings and one-way mappings studied within CMT are
inputs to and constraints on the kind of dynamic conceptual networks posited
within BT [Blending Theory].”
However, the reasons for the complementary relationship between CMT and
CIT I discuss in this section are somewhat different from theirs, in that I place
the emphasis on different aspects of the relationship within a larger, multilevel
and contextual view of metaphorical conceptualization.
Compatibility Let us begin with the issue of the role and importance
of structure mapping in CMT and CIT. Although CIT theorists urge CMT
researchers to move beyond simply studying pairwise correspondences
between two domains, it is clear that structure mapping of the kind CMT is
interested in is just as important for CIT as it is for CMT. We saw above in a
number of examples that shared structure is essential not just for metaphorical
mapping (from B to A) but also for all the other kinds of mapping and
projections in CIT that characterize metaphorical blends. The mental spaces
in double-scope networks allow for mappings not just from the source domain
to the target but also mappings from both (or more) of the input spaces into the
blended space and even from the blend into the input spaces. All of these
mappings (or projections), which are additional to the standard metaphorical
mappings (i.e., from source to target), assume compatible input spaces in the
sense that the additional mappings or projections are the result of incompat-
ibilities that themselves derive from systematic compatibilities (as captured by
systematic source-to-target mappings). This means that we cannot focus on the
incompatibilities without at the same time focusing on what is compatible
between two domains. In order to see the incompatibilities, we also have to see
the compatibilities. We cannot figure out what the incompatible elements in
blends are without identifying what the compatibilities are on which they are
built. Thinking of the relationship this way, we can suggest that CMT, on the
whole, is the study of compatibilities, whereas CIT is the study of the incompat-
ibilities that result from those compatibilities (as found in the source-to-target
6.3 Conceptual Metaphor and Conceptual Integration 147
and perform very different cognitive tasks. For this reason, we cannot expect
the large dormant system to perform all the complex cognitive tasks that the
other fully specific and fully contextualized system can. I would guess that the
simplicity of the large system is by design: to make sure that we are capable of
keeping it in long-term memory for the purposes of mobilizing it for the
intricate application of the double-scope networks in working memory.
In conclusion, rivalry is unjustified and unnecessary. We are doing different
but complementary things within a unified and many-layered system of meta-
phorical conceptualization. What CIT researchers do at the level of mental
spaces would not be possible without the findings of the new, extended CMT,
and the work in CMT done at the level of image schemas, domains, and frames
would be incomplete without the contributions of CIT.
6.4 Conclusions
The phenomenon of conceptual metaphor is simultaneously offline and online.
In the course of using conceptual metaphors, offline conceptual structures in
long-term memory (image schemas, domains, and frames) are put to cognitive
work online in mental spaces in working memory.
This view enables us to take into account a variety of metaphor-related
mental activities that speakers engage in at this level. These include getting
primed to use particular metaphors by context, giving metaphorical expres-
sions specific socio-pragmatic functions, creating novel metaphors, using
metaphors deliberately, and several others.
Two quintessential online metaphorical phenomena were discussed in some
detail in the chapter: mixing metaphors and conceptual integration. I argued
that both assume and are heavily influenced by “higher” level image-schema,
domain, and frame metaphors. The analysis of mixing metaphors and concep-
tual integration indicated that, although much of the mental activity is
happening at the level of mental spaces in both cases, it would not be possible
to account for their use without taking into account the corresponding offline
conceptual structures. In the extended view of CMT, the offline and online
metaphorical structures are all needed for a (more or less) complete under-
standing of how conceptual metaphors work in natural discourse.
7 The Shape of the Extended View of CMT
In this next to last chapter, I attempt to identify the main ingredients of the
extended view of CMT, characterize the most important distinctions it
involves, outline the broad shape of the new view, and draw out its implica-
tions for a general understanding of metaphor. In the last chapter, I revisit the
five questions that inform the content and structure of the extended view, as
presented in this book.
150
7.1 Two Major Issues for CMT 151
there was a very loud rock concert and a comparatively quiet summit meeting.
The loudness and the relative quiet of the respective events were perceptual
features of the events. In this case, the original conceptualizer (Dr. Naidoo)
was prompted to use a perceptual property of the physical context from the
large body of experiential content that was available to him.
In the beginning, I didn’t make the connection between the subject matter and my own
sero-positivity. I was asked to be part of the Day Without Art exhibition a few years ago
and didn’t think I was worthy – other artists’ work was much more HIV-specific. . . .
But my mentor said, “Don’t you see the connection? You’re documenting something
that was never intended to live this long. You never intended to live this long.” [italics
in the original]
The mentor calls Jump’s attention to the metaphorical relationship between his
illness and his art. We can state this in the following metaphorical analogy:
surviving aids despite predictions to the contrary is for the
old mural advertisements to survive their expected “life
span.” The two domains share an image schema; namely, that there is an
entity that exists longer than expected. The image schema is novel and
unconventional.
If this is on the right track, we can have three (not two) conceptual pathways
for the creation and comprehension of conceptual metaphors including meta-
phorical analogies:
schematicity hierarchy pathways
ad hoc pathways
shared image schema pathways
The third kind of pathway could serve as the basis for conceptual metaphors
that are usually regarded as metaphorical analogies. CMT research typically
foregrounds the types of metaphors that are based on schematicity hierarchies,
but there appears to be a large number of other cases that require ad hoc and
shared image schema pathways. In addition, in the section on the kinds of
metaphors in this chapter, I propose yet another conceptual pathway – that of
feature resemblance. Future research can quantify the precise proportion of the
different types.
even reject the existence of conceptual metaphors of the kind mostly discussed
in the CMT literature (like life is a journey and the primary metaphors that
give them their embodied character) and, at the same time, they emphasize the
role of context in the creation and use of metaphors.
In his paper, Gibbs criticizes the four theories of discourse metaphor and
offers a synthesis of the discourse and embodied metaphor views in an
integrated model of “metaphor performance”: the dynamic systems view of
metaphor (see, e.g., Gibbs and Cameron 2007; Gibbs 2017c). I believe that the
dynamic systems view bears many similarities to the contextual and multilevel
view of metaphor that I am developing here. Both views share the goal of
characterizing metaphor performance (just like the discourse metaphor views),
but they do so by taking advantage of (rather than rejecting) embodied
conceptual metaphors – though to different degrees. In the remainder of the
section I briefly outline how my proposed account could achieve this. To
demonstrate, I use, in addition to the metaphors in the mountain stage and
whisper, the metaphor capsize, as it was used in a piece of natural discourse in
a 2007 newspaper article I analyzed in previous work (Kövecses 2015a).
Following hurricane Katrina that devastated New Orleans in 2005, a jour-
nalist did an interview with Fats Domino, a now elderly American rock
musician living near New Orleans. The journalist writes:
The 2005 hurricane capsized Domino’s life, though he’s loath to confess any incon-
venience or misery outside of missing his social circle . . . (USA TODAY, 2007,
September 21, Section 6B) (taken from Kövecses 2015a: 103).
The interview was largely about how Domino’s life was affected by the
hurricane. Its very negative effect is expressed by the unconventional linguistic
metaphor capsize, which is an instance of the applicable conceptual metaphor:
life is a (sea) journey. At the time of the interview, there were still many
noticeable signs of the devastation, as the newspaper article about domino
indicates. The journalist continues the quote above as follows: “. . . in the
Lower Ninth Ward, still destroyed and deserted despite some signs of
renewal.”
I suggest that this salient perceptual property of the physical situation plays a
major role (but likely not the only role, as suggested to me by Michele Feist in
personal communication, March 15, 2019) in the metaphorical conceptualization
of Domino’s life: the still visible sight of the devastation caused by the hurricane
(situational context), the memory of the overturned boats (conceptual-cognitive
context), the topic of the conversation with Domino (discourse context), and the
universal experience of falling down and not being able to function (bodily
context) prime the conceptualizer (the journalist) to use the verb capsize, which
assumes (and evokes) the primary metaphors of action is self-propelled
motion and persisting or functionality is remaining erect.
156 The Shape of the Extended View of CMT
Simultaneously, since the topic of the interview is Domino’s life, the life and
travel domains are evoked, resulting in the conceptual metaphor life is
travel. An aspect of this metaphor is the correspondence between leading
one’s life and journeying, that is, the metaphor living a life is journey-
ing. (As noticed by Framenet, journeying is the activity aspect of travel,
which is a more inclusive category than journey.) And since the journalist’s
immediate idea at this point in the discourse concerns how Domino’s life has
suddenly changed for the worse, a very specific metaphor is evoked – that of
a sudden, unexpected turn of events for the worse in domi-
no’s life is the capsizing of domino’s boat in the course of his
sea journey, as expressed by the verb capsize.
In other words, primed by the context, a schematicity hierarchy is con-
structed online, which can be represented as follows:
Image schema:
action is self-propelled motion + persisting / functionality
is remaining erect
Domain:
life is travel
Frame:
living a life is journeying
Mental space:
a sudden, unexpected turn of events for the worse in
domino’s life is the capsizing of domino’s boat in the
course of his sea journey
the example of capsize. Capsizing is just a small part of the sea journey
frame. Its contextual meaning is “changing for the worse.” But it (capsizing)
has its own conceptual potential to expand by means of drawing out its
implications, such as one’s life being normal before it happens, the change
being sudden and unexpected, uncertainty about the future following the
change, the change for the worse evoking empathy and compassion, and
others. In a way, it can be suggested that a particular contextual meaning is
introduced in order to enable a variety of social, pragmatic, emotive, rhetorical,
etc., functions and effects. In traditional terminology, this is called connotative
meaning. Contextual meaning would, then, correspond to connotative mean-
ing, mentioned above.
It is this type of meaning construction that is the true “element” of concep-
tual integration; it is here that conceptual integration can break up conventional
conceptual structures, reassemble them in new ways, and can construct novel
meanings. And the same applies to deliberate metaphors that are based on
conceptual metaphors and come with a schematicity hierarchy. However, this
is not to claim, of course, that either blending or deliberate metaphors can only
rely on conceptual metaphors (those based on correlations in experience). But
in the case of conceptual metaphors based on correlations in experience, the
three kinds of meaning are jointly present in every use of a metaphorical
expression.
Here, there are multiple mappings at the schema, domain, and frame levels,
such that ELEMENT A in a source frame, domain, and schema maps onto
ELEMENT A in the target frame, domain, and schema, ELEMENT B in the
source frame, domain, and schema onto ELEMENT B in the target frame,
domain, and schema, and so on for ELEMENTS C, D, E, etc. on all three
levels. The fourth level, that of mental space, is dedicated to a single mapping,
where the metaphorical contextual meaning is conveyed.
And in the case of one class of resemblance metaphors, there is a single
mapping from the source entity to the target entity, which exploits a single
feature:
mental-space level metaphor
"
ENTITY 1: FEATURE A – ENTITY 2: FEATURE A
"
FEATURE RESEMBLANCE
An example for this case could be: “John is a pig,” where an assumed feature
of pigs (being dirty, messy) is attributed to John (by way of a metonymic
process: the whole entity for a feature that characterizes
that entity). This is essentially a situation that operates with an (assumed)
analogy (resemblance) between features of two different entities. (Moreover,
7.4 Kinds of Metaphors 161
the assumed analogy between the two entities is based on the prior personifi-
cation of pigs, that is, on the metaphor pigs (animals) are people.)
“Feature resemblance” can be seen as a further type of conceptual pathway
besides schematicity hierarchy pathways, ad hoc pathways, and shared image
schema pathways. This type of pathway produces non-systematic (isolated)
linguistic metaphors, such as pig in the example above.
But analogies are usually seen as multiple mappings between two frames or
domains. Consider the expression gut the building, as used in the following
example:
Kirk E. Kreuzwieser, vice president of Strollo Architects, says much of the interior of
the 96-year-old Wells Building needs to be gutted. He and Gregg Strollo, company
president, believe the $4 million-plus rehabilitation project could be completed in less
than nine months after financial issues are resolved. (www.vindy.com/photos/2013/jun/
17/68965/)
This is a domain- or frame-level metaphorical analogy that is constituted by
multiple mappings (Figure 7.1). The mappings are:
the person who disembowels a dead animal ➔ the person who removes the
interior of a building
the animal body ➔ the building
the bowels ➔ the interior
the action of disemboweling ➔ the action of removing the interior
The mappings are based on a structural analogy between two frames or
domains, which means that the propositionally-explicated mappings between
the two frames or domains (here, animal and building) share an image-
schematic structure, an image schema, something like this:
Source Target
ANIMAL BUILDING
Thus, structural analogy (or resemblance) means that the source and the
target share some image schema (here, internal container (or object) moved out
of external container). Of the four mappings above, the last one appears at the
mental-space level. This kind of metaphor could be represented as follows:
162 The Shape of the Extended View of CMT
In this case, unlike correlation metaphors, there is a shared image schema (not
an image-schema metaphor) and there is only a single set of mappings between
the elements – either on the frame or domain level (instead of the three sets of
mappings at the image schema, domain, and frame levels). At the mental-space
level, the mapping will be limited to one, which conveys the intended context-
ual meaning. This is the general representation for cases like gut the building
and also for cases like the Frank Jump example discussed above in the chapter.
If we employ a broader definition of conceptual metaphors as any case of
multiple mappings from a source to a target (and not just those that are
correlation-based), many metaphors that are resemblance-based (those that
map entire structures, and not just a single feature) would also qualify as
conceptual metaphors (such as the metaphor above or, say, the an atom is
the solar system metaphor). These are usually called metaphorical analo-
gies. But, when they talk about conceptual metaphors, cognitive linguists
typically mean correlation-based metaphors, and, as we saw throughout, these
come with schematicity hierarchies with four levels (as discussed above).
Situational Conceptual-Cognitive
Context Discourse Context Bodily Context Context
have a potential priming effect. A main issue I discussed in this book was how
the cognitive aspects and the contextual aspects of metaphors can be put
together in a coherent framework. This issue requires the formulation of a
hypothesis to explain how the speaker comes up with a metaphor to begin with
and how the hearer can understand this metaphor. At the beginning of the
present chapter, I looked at some examples of how these metaphor processes
can be accomplished within my extended CMT framework. No matter how
oversimplified and naïve it may be, I now attempt to turn the previous
observations about the cognitive and contextual aspects of conceptual meta-
phors into a process view, that is, into a dynamic model of mental processes in
the course of the construction of metaphorical meaning.
7.7 Conclusions
In the chapter, I have outlined the main components of, and their interrelations
in, an extended CMT. The new view builds especially on the multilevel nature
and the contextual embeddedness of conceptual metaphors.
As regards their multilevel nature, conceptual metaphors involve a variety of
components. These were identified as certain kinds of metaphorical meaning
168 The Shape of the Extended View of CMT
I structured the content of this book according to five major issues that were
presented as five questions in the titles of five chapters. These were:
(1) The abstract understood figuratively, the concrete understood literally,
but the concrete understood figuratively?
(2) Direct or indirect emergence?
(3) Domains, schemas, frames, or spaces?
(4) Conceptual or contextual?
(5) Offline or online?
These are deliberately provocative titles that express extreme and simplifying
alternative views as regards CMT. In this final chapter, I would like to briefly
discuss them one by one. It will become clear, and I hope the relevant chapters
make this evident, that I do not subscribe to any such alternative views by
themselves. Reality is much more complex than what any of the alternatives by
themselves would represent.
169
170 By Way of Conclusion: Responses to the Five Questions
the theory. If many source domains are in part figurative, then it is not the case
that abstractions are only conceptualized by means of literally understood
source concepts.
My suggestion to handle this apparent dilemma for the theory was that many
source domains have ontological content that is based on tangible experience
but they are also partially figuratively understood. Many concepts can be said
to have both ontological content and figurative construal as aspects of their
meaning. In the case of many concrete concepts the ontological content
dominates over the figurative construal aspect. In the case of abstract concepts
the reverse is the case: figurative construal dominates over ontological content
(Figure 8.1). This can be diagrammed as follows:
figurative construal
FIGURATIVE
CONSTRUAL
ONTOLOGICAL
CONTENT
ontological content
Abstract concepts have very little ontological content, but a great deal
of figurative construal (i.e., they are for the most part figuratively under-
stood). But they do have some ontological content. This is what we call
the bodily basis, the embodiment of abstract concepts (such as the body
heat of anger).
It is interesting to note in this regard that several researchers consider
emotion concepts as somewhat different from other abstract concepts (see,
e.g., Altarriba et al. 2004; Borghi et al. 2017). They suggest that emotion
concepts are more embodied than other abstract concepts. I believe, however,
that this claim is too strong and needs to be qualified. It seems to me that it is
only prototypical emotion concepts (such as anger, fear, sadness, joy) that are
more embodied (that is, have a great deal of bodily basis). Nonprototypical
ones often lack embodiment completely – at least if we limit embodiment to
external signals of emotion states. (But as Anna Borghi remarked in personal
communication, e-mail message, February 26, 2019, we should consider not
only exteroceptive signals of embodiment but also interoceptive ones, which
would further refine the picture. This is a complex issue that I cannot go into
8.2 Direct or Indirect Emergence? 171
process. I showed that this can happen in essentially two ways. (1) In an initial
“frame-like” structure with element 1 (E1) and element 2 (E2), E1 and E2 are
correlated in experience, and thus E2 can stand metonymically for E1. For
instance, body heat stands metonymically for anger. Then, element 2 is gener-
alized (or schematized) and becomes the source concept for the initial concept,
which becomes the target concept. That is to say, body heat is generalized to
heat outside the emotion domain, thus allowing the primary metaphor anger
is heat to emerge. (2) In an initial “frame-like” structure with element 1 (E1)
and element 2 (E2), E1 and E2 are correlated in experience, and thus E2 can
stand metonymically for E1. Then, E2 is generalized and becomes the target
concept for the initial concept, which becomes the source. For instance, in the
initial “going to a destination” frame going to a destination involves the
purpose of arriving at that destination (thus, the two are correlated), which
gives us the metonymy arriving at a destination for achieving a
goal. But not all goals are related to arriving at destinations. Destination-
related goals are generalized to purposes in general. This yields the purposes
are destinations primary metaphor.
It is not my intention with this analysis to suggest that primary metaphors
are in any way unimportant. On the contrary, they do play the extremely
important role in human cognition in general and metaphorical conceptual-
ization in particular that was given to them in the theory of embodied cogni-
tion. My suggestion simply is that the metaphorization process that leads to
primary metaphors is preceded by a metonymization process, a metonymic
stage. In this sense, primary metaphors do not emerge directly. The metaphor-
ization process that leads to primary metaphors is not primary; it builds on the
prior process of metonymization.
In addition to its role in primary metaphors, metonymy also plays a part in
the formation of abstract concepts in general. But while in primary metaphors
metonymy assumes two closely correlated concepts, in the formation of
abstract concepts it can perform its role within a single concept, where the
original concept involves its own schematized copy. For example, the primary
meaning of the verb gut (“to take the interior organs of a dead animal out of
it”), analyzed in a previous section of this chapter, involves as its schematized
conceptual copy “to remove something out of something.” This schematization
works by metonymy, which, in this case, would be specific meaning for
its corresponding generic meaning or, more generally, specific for
generic. Thus, in the primary physical sense we have a human agent taking
out the internal organs of a dead animal body and in the abstract sense we
have an agent removing something out of something. The generic meaning
does not specify what is taken out, how it is taken out, and what it is taken out
of. This bleached version of the primary sense of gut is a schematized copy of
8.3 Domains, Schemas, Frames, or Spaces? 173
that sense, such that the specific meaning stands metonymically for the
generic one. As a result, an abstract sense of gut is created that is applicable
to several different domains such as economy or law (as indicated, e.g., by
Merriam-Webster).
It seems then that the metonymization process is crucial not only in creating
primary metaphors but also in the emergence of abstract concepts in general. In
previous work (Kövecses 2015a: 22–24), I noted “metaphorical schematization”
works by conceptual metonymies, and the process gives us a number of abstract
concepts based on metonymy, including such concepts as heaven, provi-
dence, and causation / creation.
are inextricably bound together; one cannot exist without the other.
Conceptual metaphor theory cannot simply be just conceptual. It is just as
much contextual. The question in the section title simply does not make too
much sense.
notions of scenario, scene, and mental space as the level where metaphors are
actually used (see, e.g., Musolff 2006). Still others, like those concerned with
the pragmatic and rhetorical uses of metaphor, complained that CMT has no
way of handling the pragmatic-rhetorical functions of metaphor in natural
discourse (see, e.g., Goatly 1997; Semino 2008; Steen 2008, 2011). It grad-
ually became clear that something important was missing from CMT as a
general theory of metaphor.
The question was: How can we maintain all the achievements of CMT and
at the same time accommodate these criticisms? As I attempted to show in the
previous chapter (Chapter 6), my personal solution to this dilemma is to extend
the schematicity hierarchy to include the level of mental spaces (or scenarios).
Instead of the usual three-way distinction between image schema metaphors,
domain metaphors, and frame metaphors, I suggest a four-way distinction that
includes mental space level metaphors. The new schematicity hierarchy goes
from the most schematic to the least schematic (or most specific) level: to that
of mental spaces. Mental space level metaphors assume all the higher level
ones in a hierarchy. This way, we can turn CMT into a theory that can combine
the idea of an offline schematicity hierarchy and its online functioning in a
unified view of metaphor as both a product and a process.
As a final demonstration of how this might work, consider the example of
capsize again. As mentioned above, a journalist is interviewing Fats Domino
about his life in the wake of Hurricane Katrina. In the course of interpreting
what happened to Domino, the journalist gets to a point where he wants to say
that Domino’s life took an abrupt and bad turn. At the time of the interview all
the visible signs of the consequences of the hurricane are still present, such as
the overturned boats. He writes: “The 2005 hurricane capsized Domino’s
life,. . .” The perceptual property of what he saw then and remembers now
(overturned boats) functions as a contextual factor that activates the image
schema metaphors action is self-propelled motion, purposes are
destinations, and functionality is staying erect. The lack of
functionality (of the boat) amounts to (the boat becoming) an obstacle to
continue motion: that is, to the further image schema (primary) metaphor:
difficulties are obstacles. Since the topic of the conversation is life,
these image schema level metaphors activate the domain level metaphor life
is travel. This matches the primary metaphors and elaborates on them
(Figure 8.2). The life is travel metaphor evokes the more specific meta-
phor living a life is journeying at the frame level. Now the journalist is
in a position to take the journeying frame and to express the contextual
meaning of “abrupt negative change in Domino’s life” through the highly
specific mental space level metaphor: an abrupt negative change in
domino’s life is the capsizing of domino’s boat during his sea
journey.
8.5 Offline or Online? 179
LIFE IS TRAVEL
the exposed earth gave him the very idea to express what he did, in which case
we could suggest that it was the physical context (i.e., what he saw) that
evoked the metaphorical idea, rather than the other way around. Probably, we
all remember situations where the context gave us a metaphorical idea, rather
than the idea leading us to some context through which we expressed the
preexisting idea metaphorically.
But the view of metaphor that is most compatible with the extended view is
that of the dynamic systems view of metaphor, as proposed by Gibbs in
various publications (e.g., Gibbs 2013, 2017c; Gibbs and Cameron 2007).
Dynamic system models make use of a wealth of information as the system of
adaptive behavior moves forward. The system takes in information about the
body (both present and historical), about the conceptualizers who participate in
the communicative situation, the discourse itself, the goals of the participants,
the objects and events present in the situation, and the broader environment.
The dynamic systems view and the extended CMT share the assumption that
the totality of this information and all of our experiences can contribute to
metaphor production and comprehension. To capture this totality, the extended
CMT invokes four types of context: situational, discourse, conceptual-
cognitive, and bodily. It also assumes, probably with the dynamic systems
model, that any of this information and experience can prime the use of
metaphors as discourse progresses. Given the diversity of factors they rely
on, both views can be expected to do well in accounting for metaphorical
creativity and context-sensitivity.
Gibbs (2013: 72–73) regards metaphoric activity “as a dynamic, self-
organized process” and considers metaphor “as an emergent product of multiple
constraints operating along different time-scales. . . .” Let us return to the
example of capsize, as Gibbs (2017c: 331–332) uses this metaphor from the
article about Fats Domino to demonstrate the kinds of forces, or constraints, that
all contribute to the emergence of this metaphor. He provides the following list:
Evolutionary forces to maintain survival and enhance cooperation to
enhance individual and group welfare. Primary metaphors (e.g., persist-
ing is remaining erect) emerge from both evolutionary and experien-
tial forces.
Culturally specific ideas about different topics, such as how best to deal with
difficult life challenges.
Historical forces related to the author’s past experiences as a writer for the
specific forum in which the Fats Domino profile was published.
Social forces that enable speakers to negotiate specific meaning in particular
contexts.
Linguistic factors related to conventions about particular word and gram-
matical uses including, in this case, employing nautical terms to refer to
human experiences.
Immediately salient contextual information such as the presence of the
aftereffects of Hurricane Katrina.
Immediate communicative motivations to state something that is vivid,
memorable, contextually relevant, and succinct regarding a rather complex
abstract idea related to difficulties in one’s life.
182 By Way of Conclusion: Responses to the Five Questions
Previously stated words that primed the author’s use of certain boating and
water terms that were related to earlier talk about the physical damage to
buildings caused by Hurricane Katrina.
Bodily experiences related to enduring concepts that are relevant for speak-
ing about temporary setbacks in life, such as the idea of being unable to
persist is not being able to remain erect or upright.
Brain and neural activity including that which emerges as brain systems
become coupled during particular interaction between speakers and listeners
or writers and readers.
Let us go over this list of forces and see how the extended CMT would
handle them.
In the extended CMT, evolutionary forces play a role as the shapers of
primary metaphors, and thus are part of the bodily context. The primary
metaphors themselves, like persisting is remaining erect, are part of
the conceptual-cognitive context.
The force that Gibbs calls “culturally specific ideas about different topics”
could probably be accommodated in the system I have proposed as part of the
cultural context – if “culturally specific ideas” mean the knowledge and
experience that members of a speech community share relative to such topics.
The historical forces (in the sense of “the author’s past experiences as a
writer for the specific forum in which the Fats Domino profile was published”)
could probably be accommodated as the personal history of the author within
the conceptual-cognitive context.
The social forces “that enable speakers to negotiate specific meaning in
particular contexts” would seem to fit into what I call the social context within
the more inclusive context type of situational context.
The linguistic factor of conventionally using nautical terms to talk about life
and human experiences in general essentially boils down to how conceptual
metaphors are given linguistic expression in discourse. Here the conceptual
metaphor would be life is a sea journey. This would again be part of the
conceptual-cognitive context, and the expressions would form part of the
mental lexicon that attaches to the conceptual system.
Immediately salient contextual information corresponds to the situational
context (more specifically, the physical environment) in the extended CMT
framework. All the perceivable aftereffects of the hurricane are potential
contextual factors in the use of the metaphor.
Immediate communicative motivation (in the sense of communicative goals)
is not something that is part of any of the four context types, which deal with
actual contentful experiences. (I will say more about this issue below.)
The force of previously stated terms related to the damage caused by the
hurricane comes close to what I called “previous discourse on the same topic” –
8.6 Final Words 183
an aspect of the discourse context. This can include all the boating and water
related words that were used in the press previously.
The force of bodily experiences related to the idea of being unable to
persist is not being able to remain upright could clearly be
accounted for in the extended CMT framework – both as part of the bodily
context (as an experiential correlation) and as part of the conceptual-cognitive
context (as a variant of a primary metaphor).
Finally, the brain and neural activity can be accommodated within extended
CMT as regards the neural correlates of experiential correlations (bodily con-
text), but not as regards the neural couplings between speakers and listeners.
We can now make some further general observations concerning the simi-
larities and differences between the two approaches. As can be seen in light of
the comparison above, certain forces, or constraints, from the dynamic systems
view are not matched or cannot easily be accommodated by the extended CMT
view. This makes the dynamic systems view broader in its scope. It is
interesting to see what causes this imbalance between the two frameworks.
The major goal of the extended view of CMT so far was to capture how
“contentful experience” in various types of context can prime the use of
metaphors in particular discourse situations (see Kövecses 2015a). To this
end, four different “context types” were assumed, including the situational,
discourse, bodily, and conceptual cognitive context. These are all constituted
by various “contextual factors” that correspond to some contentful experience.
The contentful experience can be bodily but also situation-, discourse-, and
(cognitive) representation-based. This last one is what I term “conceptual-
cognitive context” and it simply means that the contentful experience is
already stored (as in the case of conventional conceptual metaphors).
If we look at the forces of the dynamic systems view in relation to capsize,
we can see that all four context types are present on Gibbs’ list. But his list
includes more, such as “the immediate communicative motivation,” that have
no obvious parallel on my list of kinds of contentful experience. The reason,
I believe, is that the missing items can be both behavior-organizing systems
and contentful experiences in the sense in which I am using the term. If we
take “the immediate communicative motivation” as an example, we can see
that communicative motivation is an organizing principle of human (communi-
cative) behavior, rather than a piece of contentful experience, such as the
capsizing of a boat (situational context), what was said about the hurricane
earlier (discourse context), or being unable to get up when seriously sick
(bodily context), or finding the primary metaphor persisting is remaining
erect in the metaphorical conceptual system (conceptual-cognitive context).
Additional organizing principles of (metaphor-related) behavior include the
purposive (telic) component of behavior, the emotion and attitude system, and
so on. For example, it can be observed that when we are angry, we tend to use
184 By Way of Conclusion: Responses to the Five Questions
space level with more or less permanent or stable conceptual structures. This
integration occurs by means of higher-level structures shaping mental spaces,
on the one hand, and mental-space level structures activating higher-level
structures, on the other. Moreover, since the mental spaces are embedded in
information-rich context, the various contextual factors also influence what
goes into the emerging mental spaces. Notice that I am not suggesting than the
extended CMT is “better” than the dynamic systems view. Actually, since my
argument assumes the existence of long-term conceptual structures, if it turns
out that there are no such conceptual structures in the mind, then clearly the
dynamic systems view is to be preferred.
In addition to the question of how the extended CMT view is related to other
theories of metaphor, there are many other pending ones. First, if the extended
CMT and dynamic systems views are compatible, how could they be turned
into a single unified theory? How can their respective advantages be best
utilized and how can their respective disadvantages be eliminated? Second,
how could either one, or the unified version, be turned into a processing model
of metaphor use? Third, how is it possible to devise experiments that work
with the totality of our experiences that the two views attempt to make use of
within their frameworks? Fourth, to continue with the same issue, how can this
bewildering complexity and diversity of information, experience, timescale,
processing, and interaction be turned into a neural theory of metaphor? Can
any computational model take all this into account, including the challenge
that comes from the contextual component? And, finally, how can we study all
this happening in human subjects?
At least to my mind, these are big questions and challenges for the future.
For the time being, however, we should probably aim lower. Let us, first of all,
see if in addition to the linguistic evidence for the schematicity hierarchy we
can find supporting evidence from other modalities. Then, let us see further
specific examples of contextual influence on how linguistic and nonlinguistic
metaphors and the corresponding hierarchies are put to use in actual cases of
metaphorical meaning making. Following that, we can begin to think about the
role and importance of context-induced hierarchies more generally in social
cognition. Given some ideas about that, we can begin to build more accurate
models for what is going on in the physical and social world, in discourse, in
the body, in the conceptual system, and in the brain, when we interact
metaphorically with each other. The “interactional turn” in the study of
metaphor is right before us in its full (?) complexity.
References
186
References 187
2010. The discourse dynamics framework for metaphor. In Cameron, Lynne, and
Maslen, Robert (eds.), Metaphor Analysis: Research Practice in Applied
Linguistics, Social Sciences and the Humanities, pp. 77–96. London: Equinox.
Casasanto, Daniel. 2009. Embodiment of abstract concepts: Good and bad in right and
left handers. Journal of Experimental Psychology: General, 138(3), 351–67.
Charteris-Black, Jonathan. 2004. Corpus Approaches to Critical Metaphor Analysis.
Basingstoke and New York: Palgrave-MacMillan.
2017. Fire Metaphors: Discourses of Awe and Authority. London-New York:
Bloomsbury.
Cienki, Alan. 2007. Frames, idealized cognitive models, and domains. In Geeraerts, D.,
and Cuyckens, H. (eds.), The Oxford Handbook of Cognitive Linguistics,
pp. 170–87. Oxford: Oxford University Press.
Cienki, Alan and Müller, Cornelia. 2008. Metaphor, gesture, and thought. In Gibbs,
Raymond (ed.), The Cambridge Handbook of Metaphor and Thought,
pp. 483–501. New York: Cambridge University Press.
Clark, Herbert. 1996. Using Language. Cambridge: Cambridge University Press.
Clausner, Tim and Croft, William. 1997. Productivity and schematicity in metaphors.
Cognitive Science, 21(3), 247–82.
Coulson, Seana. 2008. Metaphor comprehension and the brain. In Gibbs, R. (ed.),
The Cambridge Handbook of Metaphor and Thought, pp. 177–94. Cambridge:
Cambridge University Press.
Coulson, Seana and Oakley, Todd. 2000. Blending basics. Cognitive Linguistics,
11(3/4), 175–196.
2003. Metonymy and conceptual blending. In Panther, K.-U. and Thornburg, L. (eds.),
Metonymy and Pragmatic Inferencing, pp. 51–79. Amsterdam: John Benjamins.
Croft, William. 1993. The role of domains in the interpretation of metaphors and
metonymies. Cognitive Lingustics, 4, 335–70.
Dancygier, Barbara and Sweetser, Eve. 2014. Figurative Language. New York:
Cambridge University Press.
David, Oana, Lakoff, George, and Stickles, Elise. 2016. Cascades in metaphor and
grammar. Constructions and Frames, 8(2), 214–55.
Deignan, Alice. 1995. Collins Cobuild English Guides 7: Metaphor. London:
HarperCollins.
2003. Metaphorical expressions and culture: An indirect link. Metaphor and Symbol,
18(4), 255–71.
2005. Metaphor and Corpus Linguistics. Amsterdam: John Benjamins.
Dirven, René and Pörings, Ralf (eds.). 2002. Metaphor and Metonymy in Comparison
and Contrast. Berlin-New York: Mouton de Gruyter.
Edwards, Derek. 1999. Emotion discourse. Culture and Psychology, 5(3), 271–91.
El Refaie, Elisabeth. 2019. Visual metaphor and embodiment in graphic illness
narratives. New York: Oxford University Press.
Evans, Vyvyan. 2013. Language and Time: A Cognitive Linguistics Approach.
Cambridge: Cambridge University Press.
Fauconnier, Gilles. 1994. Mental Spaces. New York: Cambridge University Press.
2007. Mental spaces. In Geeraerts, D., and Cuyckens, H. (eds.), The Oxford
Handbook of Cognitive Linguistics, pp. 371–76. Oxford: Oxford University Press.
Fauconnier, Gilles and Turner, Mark. 2002. The Way We Think. New York: Basic Books.
188 References
Lakoff, George. 1987. Women, Fire, and Dangerous Things: What Categories Reveal
About the Mind. Chicago: The University of Chicago Press.
1990. The invariance hypothesis: Is abstract reason based on image schemas?
Cognitive Linguistics, 1, 39–74.
1993. The contemporary theory of metaphor. In Ortony, A. (ed.), Metaphor and
Thought. Second edition, pp. 202–51. Cambridge and New York: Cambridge
University Press.
1995. Metaphor, morality, and politics, or, why conservatives have left liberals in the
dust. www.wwcd.org/issues/Lakoff.html (First published in Social Research 62
(2).)
1996. Moral Politics: How Liberals and Conservatives Think. Chicago: The
University of Chicago Press.
2008. The neural theory of metaphor. In Gibbs, Raymond (ed.), The Cambridge
Handbook of Metaphor, pp. 17–38. Cambridge and New York: Cambridge
University Press.
Lakoff, George and Johnson, Mark. 1980. Metaphors We Live By. Chicago:
The University of Chicago Press.
1999. Philosophy in the Flesh. New York: Basic Books.
Lakoff, George and Kövecses, Zoltán. 1987. The cognitive model of anger inherent
in American English. In Holland, D., and Quinn, N. (eds.), Cultural Models in
Language and Thought, pp. 195–221. New York and Cambridge: Cambridge
University Press.
Lakoff, George and Turner, Mark. 1989. More Than Cool Reason. Chicago:
The University of Chicago Press.
Landau, Mark. 2017. Conceptual Metaphor in Social Psychology: The Poetics of
Everyday Life. New York and London: Routledge.
Langacker, Ronald. 1987. Foundations of Cognitive Grammar. Stanford: Stanford
University Press.
2008. Cognitive Grammar: A Basic Introduction. New York: Oxford University
Press.
Low, G., Todd, Z., Deignan, A. and Cameron, L. (eds). 2010. Researching and
Applying Metaphor in the Real World. Amsterdam: Benjamins.
Müller, Cornelia. 2008. Metaphors Dead and Alive, Sleeping and Waking: A Dynamic
View. Chicago: The University of Chicago Press.
Musolff, Andreas. 2001. Political imagery of Europe: A house without exit doors?
Journal of Multilingual and Multicultural Development, 21(3), 216–29.
2004. Metaphor and Political Discourse. Analogical Reasoning in Discourse About
Europe. London: Palgrave Macmillan.
2006. Metaphor scenarios in public discourse: Metaphor and Symbol, 21(1),
23–38.
2016. Political Metaphor Analysis: Discourse and Scenarios. Bloomsbury.
Narayanan, Srini. 1999. Moving right along: A computational model of metaphoric
reasoning about events. Proceedings of the National Conference on Artificial
Intelligence (AAAI '99), Orlando, Florida, July 18–22, 1999, pp. 121–28,
AAAI Press.
Pragglejaz Group. 2007. MIP: A method for identifying metaphorically used words in
discourse. Metaphor and Symbol, 22(1), 1–39.
192 References
Quinn, Naomi. 1991. The cultural basis of metaphor. In Fernandez, J. (ed.), Beyond
Metaphor: The Theory of Tropes in Anthropology. pp. 56–93. Stanford: Stanford
University Press.
Radden, Günter. 2002. How metonymic are metaphors? In Dirven, René and Pörings,
Ralf (eds.), Metaphor and Metonymy in Comparison and Contrast, pp. 407–33.
Berlin-New York: Mouton de Gruyter.
Radden, Günter and Kövecses, Zoltán, 1999. Towards a theory of metonymy. In
Panther, Uwe-Klaus, and Radden, Günter (eds.), Metonymy in Language and
Thought, 17–59. Amsterdam: John Benjamins.
Rakova, Marina. 2003. The Extent of the Literal: Metaphor, Polysemy and Theories of
Concepts. Palgrave Macmillan.
Reddy, Michael. 1979. The conduit metaphor – A case frame conflict in our language
about language. In Ortony, A. (ed.), Metaphor and Thought, pp. 284–324.
Cambridge: Cambridge University Press.
Rosch, Eleanor. 1978. Principles of categorization. In Rosch, E., and Lloyd, B. B. (eds.),
Cognition and Categorization, pp. 27–48. Hillsdale, NJ: Lawrence Erlbaum.
Ruiz de Mendoza, Francisco. 1998. On the nature of blending as a cognitive
phenomenon. Journal of Pragmatics, 30, 259–74.
Ruiz de Mendoza, Francisco and Galera, Alicia. 2014. Cognitive Modeling.
A Linguistic Perspective. Amsterdam: John Benjamins.
Ruiz de Mendoza, F. J. and Mairal, R. 2007. High-level metaphor and metonymy in
meaning construction. In Radden, G., Köpcke, K.-M., Berg, T., and Siemund P.
(eds.), Aspects of Meaning Construction, pp. 33–51. Amsterdam/Philadelphia:
John Benjamins.
Searle, John. 1993. Metaphor. In Ortony, A. (ed.), Metaphor and Thought, pp. 92–123.
Cambridge and New York: Cambridge University Press.
Semino, Elena. 2008. Metaphor and Discourse. Cambridge: Cambridge University Press.
Semino, Elena and Demjén, Zsófia (eds.). 2017. The Routledge Handbook of Metaphor
and Language. London: Routledge.
Sharifian, Farzad. 2011. Cultural Conceptualizations and Language. Theoretical
Framework and Applications. Amsterdam: Benjamins.
(ed.). 2015. The Routledge Handbook of Language and Culture. Milton Park: Routledge.
(ed.). 2017a. Advances in Cultural Linguistics. Singapore: Springer.
2017b. Cultural Linguistics. Amsterdam: John Benjamins.
Shen, Yeshayahu and Balaban, Noga. 1999. Metaphorical (in)coherence in discourse.
Discourse Processes, 28(2), 139–53.
Sinha, Chris. 2007. Cognitive linguistics, psychology, and cognitive science. In
Geeraerts, D., and Cuyckens, H. (eds.), The Oxford Handbook of Cognitive
Linguistics, pp. 1266–94. New York: Oxford University Press.
Sperber, Dan and Wilson, Deidre. 1986/1995. Relevance: Communication and
Cognition. Cambridge, MA: Blackwell.
Steen, Gerard. 2008. The paradox of metaphor: Why we need a three-dimensional
model of metaphor. Metaphor and Symbol, 23(4), 213–41.
2011. The Contemporary Theory of Metaphor – Now New and Improved! Review
of Cognitive Linguistics, 9(1):26–64.
2013. Deliberate metaphor affords conscious metaphorical thought. www.academia
.edu/363751/Deliberate_metaphor_affords_conscious_metaphorical_thought
References 193
Stefanowitsch, Anatol. 2006. Words and their metaphors. In Stefanowitsch, Anatol, and
Gries, Stefan Th. (eds.), Corpus-Based Approaches to Metaphor and Metonymy,
pp. 64–105. Berlin: Mouton de Gruyter.
Sullivan, Karen. 2006. Frame-based constraints on lexical choice in metaphor. 32nd
Annual Meeting of the Berkeley Linguistics Society (BLS), Berkeley, CA, 2006.
https://ssrn.com/abstract=1492026
2013. Frames and Constructions in Metaphoric Language. Amsterdam: John
Benjamins.
2016. Integrating constructional semantics and conceptual metaphor. Constructions
and Frames, 8(2), 141–65.
Sweetser, Eve. 1990. From Etymology to Pragmatics: Metaphorical and Cultural
Aspects of Semantic Structure. Cambridge: Cambridge University Press.
Taylor, John. 1989. Linguistic Categorization: Prototypes in Linguistic Theory.
Oxford: Clarendon Press.
Taylor, John R. and MacLaury, Robert (eds.). 1995. Language and the Cognitive
Construal of the World. Berlin: Mouton de Gruyter.
Taylor, John and Mbense, Thandi. 1998. Red dogs and rotten mealies: How Zulus talk
about anger. In Athanasiadou, A., and Tabakowska, E. (eds.), Speaking of
Emotions, pp. 191–226. Berlin: Mouton de Gruyter.
Traugott, E. C. and Dasher, R. B. 2002. Regularity in Semantic Change. Cambridge:
University of Cambridge Press.
Trim, Richard. 2007. Metaphor Networks: The Comparative Evolution of Figurative
Language. Houndmills, Basingstoke: Palgrave Macmillan.
2011. Metaphor and the Historical Evolution of Conceptual Mapping. Houndmills,
Basingstoke: Palgrave Macmillan.
Turner, Mark. 1996. The Literary Mind. New York: Oxford University Press.
Turner, Mark and Fauconnier, Gilles. 2000. Metaphor, metonymy, and binding.
In Barcelona, Antonio (ed.), Metaphor and Metonymy at the Crossroads,
pp. 133–45. Berlin: Walter de Gruyter.
Van Dijk, Teun. 2009. Society and Discourse: How Social Contexts Influence Text and
Talk. Cambridge: Cambridge University Press.
Yu, Ning. 1995. Metaphorical expressions of anger and happiness in English and
Chinese. Metaphor and Symbolic Activity, 10, 59–92.
1998. The Contemporary Theory of Metaphor in Chinese: A Perspective from
Chinese. Amsterdam: John Benjamins.
2002. Body and emotion: Body parts in Chinese expression of emotion. Pragmatics
& Cognition, 10(1), 341–67.
Zinken, Jörg. 2007. Discourse metaphors: The link between figurative language and
habitual analogies. Cognitive Linguistics, 18(3), 445–66.
Index